ION MODULATED ORGANIC TRANSISTORS - Doria

ION MODULATEDORGANIC TRANSISTORSNikolai KaihovirtaPhysicsDepartment of Natural SciencesCenter for Functional MaterialsGraduate School of Materials ResearchÅbo Akademi UniversityÅbo 2010

ION MODULATED ORGANIC TRANSISTORSNikolai KaihovirtaPhysicsDepartment of Natural SciencesCenter for Functional MaterialsGraduate School of Materials ResearchÅbo Akademi UniversityÅbo 2010

Supervisor:Prof. Ronald ÖsterbackaÅbo Akademi UniversityPre-examiners:Prof. Ludvig EdmanUmeå UniversityDr. Thomas D. AnthopoulosImperial College LondonOpponent for the public defense:Prof. C. Daniel FrisbieUniversity of MinnesotaISBN 978-952-12-2468-3Painosalama Oy, Turku, Finland 2010

Nikolai KaihovirtaPREFACEPREFACEThis thesis comprises research conducted on ion modulated organictransistors during 2006 – 2010 at the organic electronics group in theDepartment of Physics at ÅA. The reported results are achieved throughinteraction and collaboration with a number of colleagues (both withinPhysics as well as interdisciplinary), to whom I am sincerely grateful.Late in 2005, I was asked by Prof. Ronald Österbacka to start workingwith organic electronics. Ronald also became my supervisor and I amthankful for all the great discussions and the inspiration he has providedme. I was handed the independence, as well as the responsibility, tomaster my research within the theme. By this arrangement, I haveobtained qualities that I will highly appreciate in my future career as aresearcher. When I started to work in 2006, I had no previous backgroundfrom organic electronics. Dr. Tomas Bäcklund served as my mentor thefirst months, for which I am very grateful.The work together with colleagues at the Laboratory of PolymerTechnology at ÅA has been essential for many of the results in this thesis.I consider that our collaboration has been ideal: open discussions on aregular basis, innovative ideas and purposeful work towards a commontarget and by utilizing each others expertise. Therefore, I wish toacknowledge Prof. Carl-Eric Wilén, Carl-Johan Wikman and Xuehan Hefor this and also for helping me comprehend polymer chemistry.In my eyes, the articles enclosed in this thesis clearly display a positivedevelopment in my reporting skills. I have elaborated my writing bycarefully reading scientific publications written by others. However, Ihave also gained invaluable help and assistance from many of mycolleagues at the organic electronics group. I particularly wish to thankDr. Tapio Mäkelä, Dr. Himadri Majumdar, Harri Aarnio and DanielTobjörk for excellent collaboration and for stimulating discussions.The help from Dr. Kjell-Mikael Källman, our laboratory engineer, cannever be exaggerated. He is very much responsible for that our ideas andhypotheses (irrespective of how crazy they may be) can be tried out in thelab. Also, I should not omit acknowledging Dr. Kaj Höglund for all theenjoyable discussions and for the much appreciated practical advicesaround this work. Furthermore, I wish to thank all the colleagues at theDepartment of Physics for keeping up such a pleasant atmosphere.i

Nikolai KaihovirtaPREFACEOff work, my family and my friends have always been important forme. Till mina föräldrar Ingrid och Ilkka: Fastän forskningen emellanåt harverkat svår att begripa, har ert villkorslösa stöd varit oerhörtbetydelsefullt för mig.Finally, I wish to thank my Hannis for all the love and support thatyou are giving me.August 15, 2010Åboii

Nikolai KaihovirtaCONTENTSCONTENTSPREFACECONTENTSOUTLINE OF THESISLIST OF INCLUDED PUBLICATIONSAUTHOR´S CONTRIBUTIONRELATED PUBLICATIONSiiiivviiixxi1. INTRODUCTION ......................................................................................................11.1 Background............................................................................................................11.2 The OFET and the Ion Modulated Organic Transistors...................................21.2.1 Transistor Structures.....................................................................................21.2.2 Current-Voltage Analysis.............................................................................41.2.3 Three Types of Organic Transistors............................................................72. EXPERIMENTAL......................................................................................................132.1 Materials and Methods.......................................................................................132.1.1 Substrates......................................................................................................132.1.2 Contacts ........................................................................................................142.1.3 Semiconductors............................................................................................152.1.4 Insulators ......................................................................................................162.2 Measurements......................................................................................................182.2.1 Electrical Characterization .........................................................................182.2.2 Optical Characterization.............................................................................193. RESULTS AND DISCUSSIONS............................................................................213.1 Absence of Substrate Roughness Effects in Electrolyte Gated OFETs.........213.2 Self-Supported and Ion-Conducting Membranes for Low-Voltage OrganicTransistors..................................................................................................................253.3 All-Polymer Low-Voltage ECTs on Patterned Ion-Conducting Membranes.....................................................................................................................................303.4 The Role of Moisture in Hydrous Electrolyte Gated OFETs.........................343.4.1 The Effects of a Varying Ambient Humidity...........................................35iii

Nikolai KaihovirtaCONTENTS3.4.2 Irreversible Degradation at Higher Electric Potentials ..........................383.5 Improved Device Stability by Adding a Sterically Hindered PhenolicAntioxidant in P3HT ................................................................................................414. SUMMARY................................................................................................................47BIBLIOGRAPHY ..........................................................................................................49SVENSK RESUMÉ .......................................................................................................59PAPER 1 63PAPER 2 69PAPER 3 75PAPER 4 83PAPER 5 91iv

Nikolai KaihovirtaOUTLINE OF THESISOUTLINE OF THESISThe transistor is one of the greatest technological innovations of the 20thcentury. The switching and/or signal amplifying properties are utilizedin most inorganic electronic applications and a single microprocessormay contain billions of transistors. Polymeric materials providesolubility, flexibility and low cost into transistors, in particular. On theother hand, the slower switching speeds and the lower integrationdensity for organic transistors will prevent them from competing with theestablished inorganic transistors. The previous work on the low-voltage(< 3 V) hygroscopic insulator field-effect transistor (HIFET), achieved inour laboratory by T. G. Bäcklund, et. al., has served as a starting-point forthis work. The targets were to continue the research on the HIFETtowards large-scale manufacturing, to improve our knowledge on ionmodulated organic transistors (both electrolyte gated organic field-effecttransistors (OFETs) and organic electrochemical transistors (ECTs)) and todevelop novel concepts by utilizing ion-conducting membranes.The thesis is based on five papers and constitutes of two parts. Part Iis, furthermore, divided into four chapters. In Chapter 1, a comprehensivetheory on the organic transistors is provided, with particular focus on thetraditional OFET. The same theory can be directly applied on theelectrolyte gated OFETs as well, which is the central transistor type in thisthesis. The experimental details are covered in Chapter 2. Chapter 3highlights the most important findings and discussions based on the fivepapers, while Chapter 4 provides a summary of this thesis. The fivepapers are attached in Part II.The research in this thesis has been done within the Center forFunctional Materials (FUNMAT) and Graduate School of MaterialsResearch (GSMR). FUNMAT is a joint research center betweenlaboratories from Physics, Polymer Technology, Physical Chemistry andPaper Coating and Converting in Åbo Akademi University as well asPolymer Chemistry in University of Helsinki. The strategy for FUNMATis to develop materials, routes and devices towards printedfunctionalities by combining the know-how from the participatinglaboratories. FUNMAT was funded by Åbo Akademi Foundation 2006 –08 and is, from 2008 onwards, a National Center of Excellence. GSMRcombines 13 laboratories from both Åbo Akademi University andv

Nikolai KaihovirtaOUTLINE OF THESISUniversity of Turku that are involved in material research and materialtechnology.Papers 3, 4 and 5 have, furthermore, been done within the “Noveltechnology platform for mass produced inexpensive transistors onflexible substrates enabling sensing applications” (FLEXSENS) project,which is financed by the Finnish Funding Agency for Technology andInnovation (TEKES) and by some industrial partners. In FLEXSENSparticipates all the FUNMAT laboratories from Åbo Akademi Universityas well as laboratories from Oulu University, Tampere University ofTechnology and University of Turku.vi

Nikolai KaihovirtaLIST OF INCLUDED PUBLICATIONSLIST OF INCLUDED PUBLICATIONS1. Absence of substrate roughness effects on an all-printed organictransistor operating at one voltNikolai J. Kaihovirta, Daniel Tobjörk, Tapio Mäkelä, and RonaldÖsterbackaApplied Physics Letters, 93, 053302 (2008).©American Institute of Physics2. Self-Supported Ion-Conductive Membrane-Based TransistorsNikolai J. Kaihovirta, Carl-Johan Wikman, Tapio Mäkelä, Carl-Eric Wilén, and Ronald ÖsterbackaAdvanced Materials, 21, 2520-2523 (2009).©WILEY-VCH Verlag GmbH3. Printed all-polymer electrochemical transistors on patterned ionconducting membranesNikolai Kaihovirta, Tapio Mäkelä, Xuehan He, Carl-JohanWikman, Carl-Eric Wilén, and Ronald ÖsterbackaOrganic Electronics, 11, 1207-1211 (2010).©Elsevier Science B. V.4. The Effects of Moisture in a Low-Voltage Organic Field-EffectTransistor Gated with a Hydrous Solid ElectrolyteNikolai Kaihovirta, Harri Aarnio, Carl-Johan Wikman, Carl-EricWilén, and Ronald ÖsterbackaAdvanced Functional Materials, 20, 2605-2610 (2010).©WILEY-VCH Verlag GmbH5. Improved device stability by adding a sterically hinderedphenolic antioxidant in low-voltage organic field-effecttransistorsNikolai Kaihovirta, Carl-Johan Wikman, Harri Aarnio, Carl-EricWilén, and Ronald ÖsterbackaSubmitted (2010)vii

Nikolai KaihovirtaAUTHOR´S CONTRIBUTIONAUTHOR´S CONTRIBUTIONThe contribution of the author to the publications included in this thesisis the following:Paper 1. The author prepared all solutions and all the transistors thatwere fabricated by spin-coating. The author did all the electricalcharacterizations and the atomic force microscopy studies. Furthermore,the author wrote the manuscript and finalized it together with the coauthors.Paper 2. The author prepared the semiconductor solutions and fabricatedthe transistors and the electrochromic devices. The author did all theelectrical characterizations. Furthermore, the author wrote the manuscriptand finalized it together with the co-authors.Paper 3. The author prepared the solutions and designed the masks forthe flexographic printing. The author did all the electricalcharacterizations. Furthermore, the author wrote the manuscript andfinalized it together with the co-authors.Paper 4. The author prepared the semiconductor solutions and fabricatedall samples. The author did all electrical characterizations, except for theimpedance spectroscopy measurements, and contributed equally in theoptical measurements together with Harri Aarnio. Furthermore, theauthor wrote the manuscript and finalized it together with the co-authors.Paper 5. The author prepared the semiconductor solutions and fabricatedall samples. The author did the UV-Vis analysis and all electricalcharacterizations. Furthermore, the author wrote the manuscript andfinalized it together with the co-authors.ix

Nikolai KaihovirtaRELATED PUBLICATIONSRELATED PUBLICATIONS1. Low-Voltage All-Polymer TransistorsIn: D. K. Aswal, J. V. Yakhmi (Eds.), Molecular and OrganicElectronic Devices. Nova Science Publishers, Inc., Hauppauge, NY(2010) in press.2. Components and circuit arrangements including at least oneorganic field-effect transistorRonald Österbacka, Carl-Eric Wilén, Nikolai Kaihovirta, Carl-Johan Wikman, and Tapio MäkeläPatent WO 2010/010233 A13. An organic field-effect transistorRonald Österbacka, Carl-Eric Wilén, Tomas Bäcklund, and NikolaiKaihovirtaPatent WO 2008/090257 A14. A multilayer coated fiber-based substrate suitable for printedfunctionalityRoger Bollström, Anni Määttänen, Daniel Tobjörk, Petri Ihalainen,Nikolai Kaihovirta, Ronald Österbacka, Jouko Peltonen, andMartti ToivakkaOrganic Electronics, 10, 1020-1023 (2009).5. All-printed low-voltage organic transistorsDaniel Tobjörk, Nikolai J. Kaihovirta, Tapio Mäkelä, Fredrik S.Petersson, and Ronald ÖsterbackaOrganic Electronics, 9, 931-935 (2008).6. Low-Voltage Organic Transistors Fabricated Using ReverseGravure Coating on Prepatterned SubstratesNikolai J. Kaihovirta, Daniel Tobjörk, Tapio Mäkelä, and RonaldÖsterbackaAdvanced Engineering Materials, 10, 640-643 (2008).xi

Part I. SYNOPSIS

Nikolai Kaihovirta1. INTRODUCTION1. INTRODUCTIONThis chapter provides a brief motivation to organic electronics research ingeneral, and to the research on organic transistors in particular. Theessential theory on organic field-effect transistors (OFETs) is given inSections 1.2.1 and 1.2.2. Three types of organic transistors are defined inthis thesis. The differences between the transistor-types are discussed inSection 1.2.3. Parts of the theory in Chapter 1 (and in Chapter 2) have alsobeen presented in a book-chapter by N. Kaihovirta and R. Österbacka[Related Publication 1.].1.1 BackgroundEver since the first transfer resistor (transistor) operation wasdemonstrated in 1948 [1], tremendous development in the field ofelectronics has been achieved. A significant milestone for integratedcircuits technology was the invention of the metal oxide semiconductorfield-effect transistor (MOSFET) [2], which has also been important forthe organic electronics research. In 1977, the first report on a dramaticconductivity enhancement in doped poly(acetylene) was presented [3]that eventually resulted in the Nobel prize in chemistry in the year 2000.The first organic transistor functionality, based on solid-stateelectrochemical reduction and oxidation (redox) of poly(pyrrole), wasdemonstrated by White, et. al. in 1984 [4]. In 1986, Tsumura, et. al. wereone of the first groups to demonstrate an organic field-effect transistor(OFET) by using poly(thiophene) as the polymeric semiconductor [5]. Theresearch on organic transistors has been focusing on understandingmaterial properties, utilizing novel materials and on resolving challenges,such as improving the device operation efficiency as well as enhancingthe life-time and the environmental stability. Particularly π-conjugatedsemiconducting and conducting polymers have been interesting due totheir possibility of being casted from solution. These materials are alsoflexible, and in conjunction with traditional (insulating) polymers theyenable the possibility of large-scale manufacturing of organic electronicdevices at low cost. Furthermore, the compatibility of conductingpolymers with biosystems has also resulted in novel innovations, such as1

Nikolai Kaihovirta1. INTRODUCTIONelectrochemical sensors, micro-fluidic systems and drug delivery devices[6]. In other words, organic bioelectronics is an area of research thatdisplays the potential for polymeric materials.A high-performing transistor should provide a high on-current at lowvoltages. For a field-effect transistor (MOSFET and OFET) this sets arequirement to the capacitance (C) of the dielectric insulator:C κε= = , (1.1)A dC i0where C i is the effective capacitance (capacitance per unit area), κ is thedielectric constant of the insulator, ε 0 is the vacuum permittivity and d isthe insulator thickness. According to Equation (1.1), there are basicallytwo ways of increasing C i: (1) By decreasing the insulator thickness, or (2)by increasing the dielectric constant. High performing, low-voltage (< 5V) OFETs have been demonstrated with both alternatives. On the otherhand, it is challenging to reach low-voltage operation with OFETs andstill maintain robustness, stability and flexibility of the structure. Anotherroute has therefore been to replace the ordinary (dielectric) insulator withan ion-conducting insulator (e.g., an electrolyte) to construct eitherelectrolyte gated OFETs or organic electrochemical transistors (ECTs). Inthis thesis, both transistor types are defined as ion modulated organictransistors. The electrolyte gated OFET, which is the central type in thisthesis, uses the same transistor structures and follows the same currentvoltage(I-V) analysis as the conventional OFET.1.2 The OFET and the Ion Modulated OrganicTransistors1.2.1 Transistor StructuresThe OFET constitutes of thin films fabricated as subsequent layers. Thepolymeric materials are often soluble (or dispersible) in common organicsolvents, and are, therefore, easily processed by drop-casting, spincoatingand/or by various large-scale coating and printing techniques.The thickness range of each layer may vary from a few tens ofnanometers up to several micrometers. The OFET is a three-terminaldevice with a source, drain and a gate electrode. The electronic currentpath (the active channel) is formed between the source and drain at the2

Nikolai Kaihovirta1. INTRODUCTIONsemiconductor/insulator interface. The active channel is only a fewmonolayers thick. The active channel is furthermore separated from thegate by an insulator. The four basic OFET-structures are shown in Figure1.1. These are either top-contact (Figures 1.1 (a) and (b)) or bottom-contact(Figures 1.1 (c) and (d)). Furthermore, the transistor structure alsodepends on the placement of the gate. It can be either in the bottom-gate(Figures 1.1 (a) or (c)) or in the top-gate configuration (Figures 1.1 (b) and(d)).Figure 1.1. The four basic OFET-structures are (a) top-contact, bottom-gate, (b)top-contact, top-gate, (c) bottom-contact, bottom-gate, and (d) bottom-contacttop-gate.An OFET-structure should meet a few criteria. Firstly, it is crucial forthe gate to be aligned directly underneath/over the active channel, inorder to avoid stray capacitances or edge effects. Secondly, the layersapplied in a consecutive fashion demand the use of solvents that do notdissolve the underlying layer and/or increase the roughness at theinterfaces. This is particularly crucial at the active channel interface.Thirdly, the current output is dependent on the choice of devicegeometry. The top-contact, bottom-gate (Figure 1.1 (a)) and the bottomcontact,top-gate (Figure 1.1 (d)) geometries benefit from a larger injectionarea, which would result in a lower contact resistance [7]. However, theinjected charges need to travel through an undoped region of thesemiconductor in order to reach the active channel [8, 9], and thereforethe thickness of the semiconductor layer becomes a significant parameter.Furthermore, it has been shown that the injection area is notunambiguously defined by the gate overlapping area [9]. This, so called,current crowding effect, implies that the charge injection from the source3

Nikolai Kaihovirta1. INTRODUCTIONmostly occurs at the edge of the source, due to the horizontal electric fieldcreated between the source and drain.1.2.2 Current-Voltage AnalysisSince the thesis only concerns with p-channel (hole conducting)polymeric semiconductors, the following theory is applied for p-channelOFETs. However, the theory is analogous for n-channel (electronconducting) OFETs upon switching the polarity of the applied voltages.Figure 1.2.The bias-scheme for a p-channel OFET. The active channel is situatedat the semiconductor/insulator interface.For the p-channel OFET, the drain is biased negatively towards thesource, in order to inject holes from the source. The conductivity, in theactive channel, is modulated by the gate under the influence of theelectric field over the dielectric insulator (Figure 1.2). The most commonI-V characteristics are the output- (I DS-V DS) and transfer-curves (I DS-V G),also shown in Figure 1.3. In the output-curve, the drain-voltage (V DS) isswept at different gate-voltage (V G ) steps. Thus, the accumulated chargesin the active channel modulate the drain-current (I DS). The output-curvecan be divided into two regions (Figure 1.3 (a)), namely the linear (V DS

Nikolai Kaihovirta1. INTRODUCTIONwhere W and L is the channel width and length, and µ is the chargemobility. By assuming V DS 4d [10-12], where d is thethickness of the insulator. A certain degree of variation in the L-ddependence exists, for instance, due to the source and drain electrodes,the thinness of the insulator and due to the used semiconductor [13]. Forshorter channel lengths the horizontal electric field dominates and shortchanneleffects emerge. Typically, the short-channel effect results in anon-saturated I DS after pinch-off, that is, g D ≠ 0 in the saturated region.In the transfer-characteristics (Figure 1.3 (b)) V G is swept, while V DS iskept constant either in the linear region or in the saturated region. Whilethe output-curve is characterized (almost) at constant accumulation, thetransfer-curve is swept over a broader accumulation range. Therefore, thetransfer-curve provides thorough information of the switching property,stability and charge mobility of the transistor. The on/off ratio measuresthe current ratio between the on-state and the off-state. A high on/offratio is naturally desired, in order to have a distinctive on- and off-state.Based on amorphous silicon (a-Si) MOSFETs, the requirement for highperforming OFETs are on/off ratios exceeding 10 6 . The slope of thetransfer-curve, after V T , is denoted as the subthreshold slope (S). Thesubthreshold slope determines the sharpness of the transition from off- toon-state, and is defined in units of [V/decade]. A lower S implicates afaster switch. Often the turn-on voltage (V 0 ) is also extracted from the5

Nikolai Kaihovirta1. INTRODUCTIONtransfer-curve. V 0 is defined as the voltage where the transistor turnsfrom the off-state to the subthreshold region, see Figure 1.3 (b). Althoughrelated, V 0 should, however, not be confused with V T, which has a directphysical definition as described above.Figure 1.3. (a) Output- and (b) transfer-characteristics for a p-channel OFET.The transfer-curve is measured in the saturated region. The square-root of I DS isalso shown in (b).The transfer-curve can be swept both forward and back to determineany hysteresis. The hysteresis is generally a result of trapped charges atthe active channel/insulator, but also possible ionic drift, in the insulator,may cause hysteresis. Another important parameter is the chargemobility (µ). Basically, µ illustrates how fast charges move through theactive channel and can be derived from Equations (1.2) and (1.4) withrespect to V G. In the linear region this would be:∂IDS∂VG≡gm=WLμCViDS, (1.5)where g m is the transconductance. From Equation (1.5), µ can be obtained:∂I∂VDSμ =. (1.6)GLWC ViDSSimilarly for I DS in the saturated regime (Equation (1.4)):∂IDS∂VG≡ gm=WLμCi( V − V )GT, (1.7)and µ will be:6

Nikolai Kaihovirta1. INTRODUCTION∂I∂VDSμ =. (1.8)GWCiL( V − V )GAnalogously, µ can be extracted from the transfer-curve. For example, inFigure 1.3 (b), the square-root of I DS is plotted with respect to V G .According to Equation (1.4) (because the transfer-curve is measured inthe saturated region):TIDS=WCi2Lμ( V − V )GT. (1.9)A re-arrangement of the terms yields:2Lk 2μ = , (1.10)WC iwhere k is the slope of the curve, as indicated by the black line in Figure1.3 (b). The threshold voltage (V T) is obtained by extrapolating the slopeof the square-root of I DS to the intersection of the x-axis.Finally, it should be noted that µ is assumed constant at all voltages inEquations (1.2) and (1.4). However, µ varies at different V G’s for OFETs(especially in OFETs using amorphous semiconductors) due to thehopping transport in a broad Gaussian density of localized states [14].Consequently, given a mobility value, the charge concentration shouldalso be given.1.2.3 Three Types of Organic TransistorsBecause of the low power consumption and wide application area of theinorganic MOSFETs, the OFETs have naturally become the most studiedtype out of the three organic transistor types. Typically, low-voltageOFETs have either adopted various oxides as high-κ insulators [15-17],ultra-thin cross-linked polymers [10-12, 18-20] and/or self-assembledmonolayers [21-23] as low-d insulators (

Nikolai Kaihovirta1. INTRODUCTIONIt is known that doping of the polymeric semiconductor by using anelectrolyte results in a transition to metallic-like properties for thepolymer [24, 25]. In the electrolyte gated OFET the dielectric insulator isreplaced by an electrolyte “insulator”. One of the first demonstratedelectrolyte gated OFETs was the hygroscopic insulator field-effecttransistor (HIFET) [26]. As emphasized by its name, the HIFET uses apolymeric insulator that becomes ion-conducting when exposed to smallprotic solvents, such as moisture. In such an environment, the HIFEToperates as an electrolyte gated OFET.Figure 1.4. A schematic representation of the charge polarization and on how theelectric field would distribute in OFETs gated by (a) a dielectric, (b) anelectrolyte where the counter-ions are equally large, and (c) an electrolyteconstituting of cations and immobile polyanions.In the electrolyte “insulator”, a gate potential will cause anions andcations to migrate towards opposite interfaces of the electrolyte. For a p-channel transistor, the applied negative gate voltage assembles anions atthe vicinity of the electrolyte/semiconductor interface and,correspondingly, cations at the electrolyte/gate interface, as schematicallyillustrated in Figure 1.4. In the simplest model (Figure 1.4 (b)), where thecounter-ions are equally large and mobile, the ionic re-orientation createsequal electric double layers (EDLs) at respective interfaces. The electricfield in the bulk of the electrolyte is zero. Therefore, unlike the dielectricinsulator, the capacitance is independent to the thickness of the insulator[27]. Instead, the capacitance of the electrolyte is the sum of thecapacitances for both EDLs:8

Nikolai KaihovirtaCElectrolyteCCEDL11. INTRODUCTIONC+ CEDL1EDL2= . (1.11)EDL2In the simplest model C EDL1 = C EDL2 , that is, C Electrolyte = 1/2 C EDL1 . Ideally,the thickness of the EDLs is on the order of Ångströms. Hence, theresulting C i can be 10 - 100 µF/cm 2 , which exceeds the capacitance ofordinary dielectric insulators by several orders of magnitude. Unlike theOFET (Section 1.2.1), the ionic drift allows for the gate to be unalignedrelative to the active channel [28, 29]. This naturally simplifies devicefabrication. On the other hand, it also creates additional problems, forexample, concerning cross-talk between neighboring transistors.Small and mobile ions also enhance the risk for electrochemicalinteraction with the polymeric semiconductor. Most π-conjugatedpolymers are redox-active. At a threshold gate potential, anions willpenetrate the electrolyte/semiconductor interface and electrochemicallydope the semiconductor [30, 31]. Thus, the doping of the semiconductorchanges from an interfacial effect to a bulk effect. Electrochemical dopingof the polymeric semiconductor will result in a structural and electronicperturbation that can be either reversible [30, 31] or irreversible (Paper 4).In both scenarios, electrochemical doping of the polymeric semiconductorwill inevitably slow down the switching response of the transistor.However, if the transistor operation is durable and steady for a longerperiod of time, then it is less important whether the doping of thepolymeric semiconductor is governed by field-effect or electrochemistry.Furthermore, for a fast polarizable electrolyte an increase in the switchingfrequency will reduce ionic penetration into the semiconductor that willminimize electrochemical interaction [31].By utilizing an immobile anion, the problems associated withelectrochemical doping may be circumvented. The anion can becovalently bound to a polymer (polyanion) [28, 32]. In such cases, thecounter-ions are not equal in size and mobility, and the polyanionic EDLwill be distributed over a broader region, as illustrated in Figure 1.4 (c).For example, according to Equation (1.11), if C EDL2 = 10 C EDL1, thenC Electrolyte would be C Electrolyte = 10/11 C EDL1 ≈ C EDL1 . Consequently, thecapacitance of the electrolyte is dominated by its smallest term, whichwould be the capacitance over the polyanionic EDL. Indeed, lower C i´sare measured for polyelectrolytes albeit the low-voltage operation usuallyis preserved, which will be shown in Chapter 3.9

Nikolai Kaihovirta1. INTRODUCTIONPEDOT + PSS - + M + + e -PEDOT 0+ M + PSS -Scheme 1.1. The reversible redox-reaction for the conducting polymericdispersion PEDOT/PSS.Organic ECTs were developed in the mid- and late-eighties as a resultof thorough research on the redox-reaction of various π-conjugatedpolymers [4, 33-37]. Usually, the device structure is similar to the OFETstructurewith the source and drain electrodes connected by a π-conjugated polymer. The gate is furthermore separated from the activeinterface by an electrolyte. However, unlike the OFET and the electrolytegated OFET, the I-V analysis described in Section 1.2.2 does not apply forthe ECT. The electronic conductivity through the transistor channel (i.e.,the part of the π-conjugated polymer, which is covered by the electrolyte[38]) is electrochemically controlled by reducing/oxidizing the π-conjugated polymer throughout the bulk. In Scheme 1.1, for example, thereversible redox-reaction is shown for the conducting polymericdispersion poly(ethylene dioxythiophene)/polystyrene sulfonic acid(PEDOT/PSS). The difference in the electronic conductivity between thehigh-conducting PEDOT + and the low-conducting PEDOT 0 can be severalorders of magnitude [38]. Figure 1.5 illustrates an electrochemical cellcomprised by the polymeric gate, the electrolyte and the transistorchannel. The counter-ions (M + ) are provided by the electrolyte. Ifneglecting any possible electronic leakage and chemical side-reactions,then the number of charges participating in the redox-reaction can beestimated by:Q =t f∫0I Gdt , (1.13)where Q is the charge, t is the time and t f is the time at the end of theredox-reaction. Thus, the switching between the on- and off-state in theECT is time-dependent. Furthermore, the switching is governed by theionic properties of the electrolyte (bulkiness of the ions, ionic conductivityand mobility, etc) as well as by the ionic transport within the π-conjugated polymer. On the other hand, due to the redox-mechanism theorganic ECT can be built in a lateral configuration (apart from the verticalgeometries shown in Figure 1.1). Furthermore, the device is insensitive tosurface roughness [38], especially compared to OFETs. It is noteworthythat the gate electrode (as well as the source and drain electrodes) canalso be metallic as long as the transistor channel is redox-active [4, 33-37].Organic ECTs have particularly been studied for sensory applications,10

Nikolai Kaihovirta1. INTRODUCTIONdue to their ability of transducing an ionic signal to an electronic signal[39-41].Figure 1.5. The electrochemical cell can be used to describe the redox-reactionbetween the transistor channel/electrolyte and the gate/electrolyte.11

Nikolai Kaihovirta2. EXPERIMENTAL2. EXPERIMENTALThis chapter reviews the polymeric materials and experimentalprocedures that were routinely used throughout this thesis. The mostessential characterization techniques are described in Section 2.2.2.1 Materials and Methods2.1.1 SubstratesHigh performing OFETs are usually manufactured on rigid and smoothsubstrates such as silicon. However, in order to fully exploit the benefitsof polymeric materials, the choice of inexpensive and flexible substratessuch as plastic and paper must be considered. These substrates aresuitable for roll-to-roll fabrication, and the transparency of the plasticsubstrate is beneficial for novel applications. On the other hand, theplastic and paper substrates provide challenges, for instance, withincreased physical roughness, liquid sorption and poor chemicalresistance. These obstacles may be overcome by applying a planarizinglayer and/or a barrier layer [42]. Another challenge is the low processingtemperatures (< 300 °C) of plastic and paper substrates that efficientlyhinders fabrication of inorganic transistors onto these substrates.The polyester-type substrate poly(ethylene terephtalate) (PET)-film(Chart 2.1 a)) had previously been used for fabricating HIFETs [26, 27]. Itwas, therefore, a natural choice as a plastic substrate for continuing worktowards the all-printed HIFET (Paper 1). In the HIFET, the only purposefor the plastic substrate is to support the transistor structure. The mostfrequently used PET-films were PET 505 and Mylar A (Dupont TeijinFilms). PET 505 has a smoother surface than Mylar A. On the other hand,the transfer of polymeric solutions by coating techniques was improvedby using Mylar A. Therefore, it was selected as plastic substrate for thecoating experiments. The PET-films were carefully cleaned by de-ionizedwater (DI water), acetone and isopropanol prior to use.13

Nikolai Kaihovirta2. EXPERIMENTAL(a)OOC COOnH FnChart 2.1. The chemical structures for (a) PET and (b) PVDF.Unlike the HIFET, the substrate in the membrane based organictransistors (Papers 2, 3, 4 and 5) acts both as support and as electrolyte“insulator”. A commonly used base-material for the membrane was thepartly fluorinated poly(vinylidene fluoride) (PVDF) (Chart 2.1 b)). PVDFis known for its durability and chemical resistance and is thereforesuitable for further manufacturing of devices. Because of the electrolytewithin the membrane matrix, no extensive substrate-cleaning procedureswere performed on the membranes. In Paper 3, however, the membraneswere boiled in DI water before further processing in order to re-activatethe hygroscopic electrolyte in the membrane and, hence, improve theionic conductivity.(b)HF2.1.2 ContactsA characteristic feature for the transistors in this thesis (with exception forPaper 3) is the use of conducting polymers for the gate electrode andmetals (bulk and nanoparticle) for the source and drain electrodes. This issimply due to the different requirements set for the electrodes in theelectrolyte gated OFETs as explained below.In Section 1.2.3 it was discussed that, for an electrolyte gated OFET, thealignment of the gate relative to the active channel is not critical.Therefore, drop-casting of a conducting polymeric dispersion providessufficient accuracy. Moreover, any risk of deteriorating the underlyingpolymeric layer can be avoided since drop-casting is a gentle method andthe polymer does not require any heat treatment in order to becomeconducting. Therefore, the gate was applied by drop-casting a conductingpolymeric dispersion in air (with exception for the all-printed HIFET inPaper 1). For drop-casting, either the water based PEDOT/PSS (BaytronP, H.C. Starck Gmbh) or the highly conducting salt of the emeraldinebase-state of polyaniline (PANI) (Panipol T, Panipol Oy) were used.Chemical structures of the two conducting polymers are shown in Chart2.2. For the all-printed HIFET, the gate was inkjetted in air with a DimatixMaterials Printer (DMP-2800, FUJIFILM Dimatix, Inc.). The inkjettable14

Nikolai Kaihovirta2. EXPERIMENTALand water based Baytron Jet PE FL and Baytron Jet HC (H.C. StarckGmbh) PEDOT/PSS dispersions were used. The major concerns inselecting the proper conducting polymeric dispersions are the risk fordissolving or reacting with the underlying layer. Some of the usedmembranes, in Papers 2, 4 and 5, heavily reacts with the water basedPEDOT/PSS, resulting in considerable swelling of the membrane matrix.In such cases the toluene based PANI (Panipol T) was a better choice.(a)(b)A -O ONHnNH +NHNH +SnnA -- +SO 3 HChart 2.2. The chemical structures for (a) PEDOT/PSS and (b) the emeraldinesalt form of PANI.In order to maintain consistency of the electronic behavior betweentransistors, it is essential to carefully control the dimensions of the activechannel (L and W). In this thesis, vacuum evaporation of gold through ashadow mask was used in order to yield the source and drain electrodes.Gold provides a nearly ohmic contact towards the used p-channelsemiconductor. The contact resistance is further minimized by the highconductivity of metals. Vacuum evaporation also allows for the thicknessof the structures to be controlled in a nanometer-scale resolution. For theall-printed HIFETs (Paper 1) the source and drain were inkjetted in airwith DMP-2800. A silver nanoparticle dispersion (Cabot Corp.) was usedfor the purpose.2.1.3 SemiconductorsPolymeric semiconductors benefit by being soluble (or dispersible) incommon organic solvents and processable at low temperatures. Thus,they are attractive for roll-to-roll manufacturing purposes. In literature, awell-known polymeric semiconductor is the regioregular poly(3-hexylthiophene) (P3HT) (Chart 2.3) [43-45]. Because of the hexyl sidechains,the P3HT is soluble in common organic solvents. The hexyl sidechainis also short enough for good electronic transport. Moreover, byattaching the side-chains in a regioregular pattern, steric twisting of thepolymer is reduced and the polymer self-organizes into a lamellar15

Nikolai Kaihovirta2. EXPERIMENTALstructure [46-48]. For OFETs, the measured mobility and on/off ratio ofwell-ordered P3HT has been reported higher than 0.1 cm 2 /Vs and 10 6respectively [49-51], which are comparable to the inorganic a-SiMOSFETs.C 6 H 13SnS C 6 H 13Chart 2.3. The chemical structure for the (regioregular) P3HT.In this thesis, the P3HT semiconductor was used for all transistors(with exception for the all-polymer membrane based ECT in Paper 3).Two P3HT materials (Rieke Metals, Inc. and Plexcore, Inc.), supplied bySigma-Aldrich Inc., was used. The P3HT was dissolved in eitheranhydrous chloroform (when spin-coated) or in a 1:1 mixture of tolueneand p-xylene (when coated by reverse gravure). The P3HT was spincoated,except for the all-printed HIFET in Paper 1. For the all-printedHIFET, the P3HT was applied in air by the reverse gravure coatingtechnique. The reverse gravure coater was a tabletop Mini-Labo testcoater (Yasui Seiki Co.).For the membrane based ECTs in Paper 3, the redox-active polymerwas PEDOT/PSS (Baytron P). It was printed by flexography using aFlexiproof 100 (R K Print Coat Instruments Ltd) flexographic printingunit.2.1.4 InsulatorsPolymeric dielectric insulators are soluble and often processable at lowtemperatures and, hence, suitable for plastic electronics. Moreover, unlikethe frequently used SiO 2-insulator, all four transistor geometries can beemployed (Figure 1.1). Polymeric insulators are, however, low-κinsulators and their break-down voltages are usually low. Therefore, andas discussed in Section 1.2.3, it has been popular to add a cross-linkingagent in order to fabricate “sub-100 nanometer” robust polymericinsulators for low-voltage OFETs. On the other hand, the problem withsubstrate roughness becomes more significant. In Paper 1, a noncrosslinkedpoly(methyl methacrylate) (PMMA) (Sigma-Aldrich, Inc.)dielectric insulator was used for fabricating OFETs on different plasticsubstrates. The chemical structure is shown in Chart 2.4 a). The PMMA16

Nikolai Kaihovirta2. EXPERIMENTALwas dissolved in ethylacetate and spin-coated onto the P3HTsemiconductor.(a)(b)CH 3OOnnCH 3OH(c)(d)CF 2 CF 2CF 2 CFCFnCH CF 2CH 2m2mO CFCF 2 CF 3H 2 CCHnO ooCF 2F 2 CSO 3 HSO 3 HChart 2.4. The chemical structures for (a) PMMA, (b) PVP, (c) PTFE:PFSA and(c) PVDF:PSSA.The electrolyte “insulators”, which were described in Section 1.2.3, arethe central focus in this thesis. The ones used in this thesis arehygroscopic, that is, they require moisture to some extent for sufficientionic drift. For the HIFETs in Paper 1, a hygroscopic poly(vinyl phenol)(PVP) was used (Chart 2.4 (b)). PVP is also a well-known polymericdielectric insulator. In particular, by cross-linking the PVP the insulatingproperties are excellent [11, 18-20]. By exposing the non-crosslinked PVPto small protic solvents the PVP becomes proton conducting. The effect isutilized for operating the HIFETs at 1 V in air [26, 27]. In this thesis, thePVP (DuPont) was dissolved in isopropanol. It was either spin-coated orreverse gravure coated (for the all-printed HIFET) on the P3HT. The drythickness of the PVP-insulator was between 1.0 - 1.5 µm.All membranes in Papers 2, 3, 4, and 5 have a thickness between 50 -150 µm. The two ion-conducting membranes that have received mostattention in this thesis are tetrafluoroethylene-perfluoro-3,6-dioxa-4-methyl-7-octenesulfonic acid copolymer (PTFE:PFSA) andpoly(vinylidene) poly(styrene sulfonic acid) (PVDF:PSSA) (Charts 2.4 (c)and (d)). PTFE:PFSA (trade name Nafion 115 (DuPont)) is a commercially17

Nikolai Kaihovirta2. EXPERIMENTALavailable fully fluorinated membrane that was purchased fromElectroChem Inc. It was used as received. The PVDF:PSSA (and the otherion-conducting membranes presented in Paper 2) was prepared by theelectron beam irradiation induced grafting (EB) technique [52, 53]. Thepreparation of the PVDF:PSSA is illustrated in Scheme 2.1. Basically theEB technique consist of three major steps: (1) electron beam irradiation,(2) grafting, and (3) sulfonation (functionalization). The PVDF-basematerial was purchased from Goodfellow, Inc. and was washed withchloroform before irradiation.i) e - -beam irradiationii) styrene/isopropanol, 80 o CCHCH 2CF22 CF2nmi) 0.5 M chlorosulfonicacid/1,2-dichloroethane, 30 o CCH HC2 CF2 CH HC2 CF2CF2CF2n mn mCH 2CHCH 2 CHppScheme 2.1. The preparation scheme for the PVDF:PSSA. The three major stepsare (1) irradiation, (2) grafting and (3) sulfonation (functionalization).SO 3 H2.2 Measurements2.2.1 Electrical CharacterizationMost I-V and all current-time (I-t) analyses were done with an Agilent4142B Parameter Analyzer (Agilent Tecnologies, Inc.) connected to acomputer. While the OFETs, in Paper 1, were characterized in a nitrogenglovebox, the HIFETs were measured in air within a relative humidity(RH) interval of 25 – 35 %. In Paper 2 all MemFETs, except for thepatterned PVDF:PSSA-based MemFET were measured in a nitrogenglovebox. The patterned MemFET was characterized in ambient air. InPaper 3, the I-V measurements were performed in a controlled humidityglovebox using a Keithley 2636 System Source Meter (source and drain)and a Keithley 2400 Source Meter (gate) (Keithley Instruments, Inc), bothconnected to a computer. While the PTFE:PFSA-based MemFETs, inPaper 4, were characterized both in a nitrogen glovebox and in RH = 5 –10 %, they were solely measured in a nitrogen glovebox in Paper 5. Allinstruments were controlled with suitable LabView software (NationalInstruments Corp.).18

Nikolai Kaihovirta2. EXPERIMENTALIn Paper 4, the I-V characterization of the PTFE:PFSA, at constantpolarity (DC), was done both in ambient air and in a nitrogen glovebox.25 nm thick gold electrodes were evaporated in vacuum, through ashadow mask. For the measurements through the membrane, theelectrode area was set to 0.5 cm 2 , while on the membrane the channelwidth was 1.0 cm and the channel length 1.0 mm.For the impedance spectroscopy measurements in Paper 4,approximately 30 nm thick gold electrodes (surface area 0.5 cm 2 ) wereevaporated in vacuum on both sides of the membrane. The sample wasplaced in an aluminum chamber and the chamber was vacuum pumpedto a level of 10 -3 – 10 -4 mbar in which the measurements were done. Theimpedance spectroscopy measurements were carried out using a GamryReference 600 potentiostat (Gamry Instruments). The potentiostat wasconnected to a computer and controlled by suitable software. The ACamplitude was 15 mV and the frequency ranged from 1 MHz to 1 Hzwith 10 points per decade. C i was calculated from the obtained reactancevalues.2.2.2 Optical CharacterizationIn Paper 5, the optical UV-Vis absorption analyses were performed in airwith a Hitachi U-3200 Spectrophotometer (Hitachi High-TechnologiesCorp.) connected to a computer. The step length was set to 2 nm. Duringthe thermal stress the samples were removed from the hotplate andallowed to cool down to room temperature prior to scanning.In Paper 4, the optical transmission through thegold/PTFE:PFSA/P3HT/gold structures were measured in air. Anincandescent tungsten lamp with a 435 nm cutoff filter was used as probelight. The light was focused on a 0.5 × 0.5 cm 2 area on the contact of thedevice, and the transmitted light was collected and focused into aSpectraPro 300i monochromator (Acton Research Corp.). Themonochromated light was then measured using a Si-detector, apreamplifier and a SR830 lock-in amplifier (Stanford Research Systems,Inc.). The instruments were controlled with LabView software. Forspectral measurements a wavelength range of 450 – 850 nm was used,and this range was scanned for applied voltages ranging from +2 V to -3Von the gate electrode. The voltage was applied for three minutes in orderto allow the sample to reach equilibrium. In Paper 4, the transmission wasmonitored at the optical absorption region by continuously scanningfrom 540 nm to 550 nm with no pauses and simultaneously applying asequence of voltages.19

Nikolai Kaihovirta2. EXPERIMENTALThe atomic force microscopy (AFM) images in Paper 1 were scannedin ambient air and in non-contact mode with an AutoProbe CP (ParkScientific Instruments). All material preparation and experimental stepswere handled in a nitrogen glovebox unless otherwise mentioned.20

Nikolai Kaihovirta3. RESULTS AND DISCUSSIONS3. RESULTS AND DISCUSSIONSThis chapter is divided into five sections. Each section presents the mainresults and conclusions for respective paper that is covered in this thesis.The all-printed HIFET is presented in Section 3.1, with particular focus onthe roughness of the used plastic substrate (Paper 1). The concept ofutilizing ion-conducting membranes for manufacturing electrolyte gatedOFETs (MemFETs) is demonstrated in Section 3.2 (Paper 2). In Section 3.3,the ion-conducting membranes are further introduced in all-polymerECTs (Paper 3). Section 3.4 emphasizes the various deteriorating effectscaused by the hydrous solid electrolytes on the OFET-behavior (Paper 4).Moreover, the idea of using sterically hindered phenolic antioxidants inorder to inhibit device degradation is motivated in Section 3.5 (Paper 5).3.1 Absence of Substrate Roughness Effects inElectrolyte Gated OFETsThe work evolved from the experiments on reverse gravure coating ofboth P3HT and PVP onto plastic substrates [54]. It was noted that arougher plastic substrate (Mylar A, root mean square (RMS) roughness =25 – 50 nm) was better suited than the smoother PET 505 (RMS roughness= 4 – 6 nm), for coating the second layer (PVP) without removing anddeteriorating the first layer (P3HT). For a conventional OFET, anincreased interfacial roughness will significantly alter the electronictransport in the active channel [51, 55-58]. However, for the all-printedHIFET, the I-V analysis did not reveal such defects albeit the activechannel became rougher (Figure 3.1). Consequently, the target for thiswork was to point at the difference between the electrolyte gated OFET(HIFET) and the OFET when changing the plastic substrate from thesmoother PET 505 to the rougher Mylar A. Additionally, a few possibleexplanations to the observed differences were discussed.21

Nikolai Kaihovirta(a)3. RESULTS AND DISCUSSIONSChannelSource & drain790 nm-I d[A]85 µm2.01.5(b) V g=-0.8V1.0-0.6 V0.5-0.4 V-0.2 V0.00V0.0 0.5 1.0-V d[V]abs(I d)[A]10 -510 -610 -785 µmV d=-1.0V-3 -2 -1 0 1V g[V]Figure 3.1. (a) The AFM picture illustrates the roughness of the active channel(after coating the Mylar A with P3HT) in the all-printed HIFET. The (b) outputand(c) transfer-characteristics of the all-printed HIFET. No pronounceddeterioration due to increased roughness is observed.(c)A comparison between the OFET and the HIFET necessitates thefabrication of the devices by spin-coating and vacuum evaporation ratherthan using the coating and printing techniques. Hence, externaluncertainties from, for example, impurities and variations in structuraldimension are avoided. Therefore, the devices were manufactured bysimilar materials and methods and as much as possible in an inertatmosphere. A bottom-contact top-gate device structure was used forboth the OFET and the HIFET (Figure 1.1 (d)). The fundamentaldifference, of course, was the use of the PMMA dielectric for the OFETand the hygroscopic PVP insulator for the HIFET. The thicknesses forrespective insulator were estimated to 0.8 µm for the PMMA and 1.5 µmfor the PVP. Eight OFETs and 14 HIFETs were studied for bothsubstrates. The lower number of OFETs was enough to verify the effect ofthe rougher substrate, as previously observed in literature [56, 57]. Sincethe effect of the increased substrate roughness on the HIFET operationwas particularly interesting in this study, a higher number of HIFETswere studied. Furthermore, a higher yield of HIFETs on both substratessimplified the statistical approach.Transfer-curves for each device and on each substrate are shown inFigures 3.2 (c) and (d). All transfer-curves were measured in the saturatedregion. Additionally, the square-root of the drain currents are presentedin Figures 3.2 (e) and (f). While the HIFET seems to be unaffected by the22

Nikolai Kaihovirta3. RESULTS AND DISCUSSIONSincreased roughness, the OFET displays expected deterioration, such as alower on/off ratio, a lower subthreshold slope (S) and a decreasedmobility. The change in mobility (µ) is interpreted from the slope (k) ofthe curves in Figures 3.2 (e) and (f). According to Equation (1.9), andwhen assuming that W, L, C i and (V G - V T) are constant, the change in k isgoverned by the change in µ.(a)48 nm(b)280 nmabs(I d)[A]sqrt(I d)[10 -3 A 1/2 ]10 -510 -610 -710 -82.01.51.00.520 µmV d=-1.5VPET 505Mylar A20 µm(c)PVP(e)PVPV d=-1.5V0.0-3 -2 -1 0 1V g[V]abs(I d)[A]sqrt(I d)[10 -3 A 1/2 ]10 -810 -910 -1010 -1110 -120.20.120 µmV d=-25V20 µm(d)PMMA(f)0.0-30 -20 -10 0 10 20V g[V]PMMAV d=-25VFigure 3.2. AFM pictures of (a) PET 505 and (b) Mylar A. Typical transfercharacteristicsfor (c) HIFETs and (d) OFETs manufactured on both substrates.The transfer-characteristics are measured in the saturated region. The squarerootof I DS for (e) HIFETs and (f) OFETs.Table 3.1 summarizes the statistical results for both the OFET and theHIFET. These results support the observations made in Figure 3.2. Alsothe yields of functioning devices (i.e., devices that are comparable to thetypical I-V curves in Figure 3.2) are shown. For the OFETs, the yieldreflects the change in the substrate roughness. A higher RMS roughnessincreases the risk for electrical breakdown for the PMMA dielectric. ThePVP is, however, almost twice the thickness of the PMMA. Moreover, theHIFET operates at substantially lower gate voltages. These two factorsaffect the higher yield of HIFETs on both substrates. The yield remainsunaffected by the increasing substrate roughness. Another indicator, infavor for the HIFET, is the calculated standard errors. For the OFET, the23

Nikolai Kaihovirta3. RESULTS AND DISCUSSIONSstandard errors reveal larger fluctuations in quality between the OFETs,while the HIFETs show more consistency between each other.Device Substrate On/off S[V/decade]k[A 1/2 /V]Yield[%]HIFET PET 505 10 2 0.80±0.05 -(9.2±0.3)×10 -4 > 70HIFET Mylar® A 10 2 0.70±0.05 -(9.1±0.3)×10 -4 > 70OFET PET 505 10 3 -10 4 3.9±0.3 -(7.4±1.2)×10 -6 ~ 40OFET Mylar® A 10 2 -10 3 7.0±0.8 -(4.9±0.6)×10 -6 ~ 20Table 3.1. Average values of the on/off ratio, the subthreshold slope (S) and theslope of the square-root of I DS (k) for all devices on respective substrate. Thestandard error has been calculated for S and k. Also the yields of functioningdevices are shown.Two arguments are pointed out to explain the higher stability inHIFETs. Firstly, it has earlier been shown that the HIFET is governed bythe potential difference rather than by the electric field [27]. The ionmigration at applied gate voltages causes the operation of the HIFET (aswell as other electrolyte gated OFETs) to be independent to the insulatorthickness. Thus, the thickness of the PVP layer does not alter the lowvoltagebehavior. Likewise, a similar smoothening effect could bepossible at the rough semiconductor/insulator interface. Secondly, thehigher capacitance that is gained by the hygroscopic PVP (as discussed inSection 1.2.3) generates more charges to the active channel to efficientlyfill all the deeper trap states.Besides the demonstrated all-printed HIFET on the Mylar A substrate[59], this work inspired to proceed towards even more challengingsubstrates, such as paper. Indeed, our group managed to demonstrateone of the first low-voltage polymeric transistors manufactured on apaper substrate [42]. After the publication of this work, other groups haveindependently also acknowledged the electrolyte gated OFET as beinginsensitive to an increased roughness [60, 61].24

Nikolai Kaihovirta3. RESULTS AND DISCUSSIONS3.2 Self-Supported and Ion-Conducting Membranes forLow-Voltage Organic TransistorsIon-conducting polymeric membranes (i.e., solid electrolytes) have beenused in liquid and gas separation, fuel-cells and in the chlor-alkaliindustry. The thicknesses of the membranes are in the order of tens orhundreds of micrometers, which enables them to be self-supportive. Bycombining the bifunctional property of the membrane (the ionconductingand self-supporting properties) with the electronic propertiesof the π-conjugated polymers, various and innovative organic electronicapplications could be envisioned. The facility and know-how ofpreparing ion-conducting membranes were provided by the Laboratoryof Polymer Technology. Therefore, the aim of this work was to present aconcept of using ion-conducting membranes in organic electronics, andparticularly, as a solid electrolyte “insulator” in OFETs (MemFETs).Figure 3.3. (a) The device structure for the MemFET. The (b) output- and (c)transfer-characteristics for the PTFE:PFSA-based MemFET. (d) The switchon/switch-offresponse for the PTFE:PFSA-based MemFET.The commercially available PTFE:PFSA (Nafion 115) is used as the“proof-of-concept” membrane (Figure 2.4 (c)). The membrane istransparent, has excellent mechanical properties and provides stable ionicconductivity at controlled atmospheres. Its morphology and ionic25

Nikolai Kaihovirta3. RESULTS AND DISCUSSIONStransport properties have been thoroughly studied [62]. The hydrophobicfully fluorinated PTFE-backbone provides the support and the chemicalresistance. In conjunction with the hydrophilic sulfonic acid groups thePTFE:PFSA self-organize to form hydrophilic nano-channels for ionicmigration [63]. The self-organized morphology also results in a selectiveionic transport. While protons migrate through the channels, anionic driftis prevented. The ionic conductivity is, furthermore, dependent on thewater content in the membrane. Upon hydration, the channels swell andthe ionic migration is improved, and vice versa upon drying.The MemFET structure as well as the I-V and I-t characteristics for aPTFE:PFSA-based MemFET are shown in Figure 3.3. Apart from being amechanical support, the membrane also acts as a platform for applyinglayers both on top and underneath the film. The characteristics aremeasured after one day of storing the devices in a dry atmosphere. Theon/off ratio for the PTFE:PFSA-based MemFET is 10 2 – 10 3 . The turn-onvoltage (V 0) has a slight positive offset (0.3 V), much similar to the HIFET(see Section 3.1, Paper 1). Earlier reports suggest that P3HT isunintentionally doped by moisture in the insulator [27]. In Figure 3.3 (d),the switch-on reaches 65 % of the maximum on-current within onesecond. The switching time for the PTFE:PFSA-based MemFET can befaster by improving the ionic motion [64]. It will be shown in Section 3.4(Paper 4) that the hydration level in the membrane largely affects thetransistor-characteristics and presumably also the switching response. Ahydrous PTFE:PFSA will improve the proton mobility and, hence, lowerthe polarization time.For large-scale manufacturing purposes, the preparation of ionconductingmembranes by the EB technique provides a number ofbeneficial properties. The EB technique consists of three major steps(Scheme 2.1). Various polymeric base-materials can be used for electronbeam irradiation. For all model membranes in Figure 3.4, the partiallyfluorinated PVDF was used, due to its similar properties compared to thePTFE. Additionally, the preparation of the PVDF is cheaper than thePTFE, which is a significant argument for large-scale manufacturingpurposes. Sulfonic acid is the functional group (PVDF:PSSA) in Figure 3.4(a), while, in Figure 3.4 (b), aminated vinylbenzyl chloride is grafted tothe PVDF (PVDF:aminated VBC). In Figure 3.4 (c) the PVDF:PS is hostinga liquid lithium salt-solution of lithium hexafluorophosphate (LiPF 6) andaprotic solvents.26

Nikolai Kaihovirta3. RESULTS AND DISCUSSIONSFigure 3.4. The chemical structures and the output-characteristics for MemFETsgated with (a) PVDF:PSSA, (b) PVDF:PS aminated VBC and (c) PVDF:PSimmersed with a liquid solution of LiPF 6 and aprotic solvents.The output-behavior for each model membrane differs drastically fromeach other which could reveal different doping conditions [65]. Thedifferent functional groups cause different ions to migrate in themembranes. The similarities between the output-curves of the PTFE:PFSAand the PVDF:PSSA is anticipated, since both membranes have the anionscovalently bound to the polymeric backbone and are, thus, immobile. InFigures 3.3 (b) and 3.4 (a), I DS is modulated by the electrostatic field-effectcaused by the polyanionic EDL. On the other hand, both thePVDF:aminated VBC and the PVDF:PS immersed with a liquid solutionof LiPF 6 and aprotic solvents contains mobile anions (Cl - and PF 6-respectively). Upon applying a negative V G the anions migrate towardsthe membrane/P3HT interface. Depending on the bulkiness of the anion,the permeability of the P3HT and the applied V G the degree ofelectrochemical doping the P3HT may vary [30, 65]. Unintentionalelectrochemical interaction of the P3HT is preferably avoided, since ionicdrift into the polymeric semiconductor slows down the switchon/switch-offof the device. Electrochemical bulk doping of thepolymeric semiconductor may also result in a weakened saturation of I DSor in a completely lost pinch-off. Moreover, the P3HT is structurally andelectronically perturbed, which may lead to lower charge mobility,degrading performance and a reduced device lifetime. Some of theseissues are further clarified in Section 3.4 (Paper 4). On the other hand,intentional electrochemical doping of the π-conjugated polymer can alsobe utilized for transistor-applications, as discussed in Section 1.2.3. Themembrane-based ECT is demonstrated in Section 3.3 (Paper 3).The key advantages for the ion-conducting membrane are thesupporting structure and the ability to utilize the ionic transport for27

Nikolai Kaihovirta3. RESULTS AND DISCUSSIONSseveral devices on the same membrane. For the latter, however,patterning of the electrolyte in the membrane would often be necessary inorder to avoid unintentional crosstalk between neighboring devices.Furthermore, grafting of different functional groups on isolated parts ofthe same membrane could be envisioned for sensing purposes and forvarious devices. In the EB process, patterning is simply achieved bymasking the base material during electron beam irradiation.Consequently, the grafting and the functionalization occur only at theunshielded (irradiated) parts of the base material. In this work, the PVDFwas shielded by a patterned aluminum mask, to form PVDF:PSSA-dots (4mm in diameter) separated by 10 mm from each other in a regular pattern(Figure 3.5). MemFETs were built both on top of the PVDF:PSSA-dots aswell as on the unfunctionalized area to demonstrate the effect ofpatterning. It is noteworthy that, unlike the ordinary PVDF:PSSA-basedMemFETs, it was possible to characterize the patterned MemFETs in airwithout any larger gate leakage currents caused by the ambient humidity.This can be explained by the restricted ion-conducting area, whichprevents unconstrained diffusion of moisture in ambient air. Moreover,the PVDF:PSSA-dot is covered on both sides by the subsequent layers inthe MemFET-structure.a)10.0b)V G=-1.2V10 mm-Drain current/ A8.06.0-0.8 V4.0-0.4 V2.00V0.0V Gfrom 0 V to -1.2 V0.0 0.4 0.8 1.2-Drain voltage/ VFigure 3.5. (a) The PVDF with ion-conducting PVDF:PSSA-dots in a regularpattern. (b) The MemFET on a PVDF:PSSA-dot (filled symbols) displaystransistor behavior, whereas the MemFET on the bare PVDF only shows ohmicbehavior (open symbols).In this work, a MemFET and an electrochromic (EC) display pixel werefabricated on the same membrane (Figure 3.6) in order to demonstrate theability of using the ionic properties for several and different organicelectronic devices. Electrochromism is a well-known feature in π-conjugated polymers [66]. The reversible and the bias controlled colorswitchusually follows two routes; either a switch between a colored anda bleached state or between two colored states. In the EC display pixel,28

Nikolai Kaihovirta3. RESULTS AND DISCUSSIONSthe electrolyte is used for inducing polaronic states by reversible redoxreactionsand thereby achieving the switch in color. The emeraldine saltform of PANI is reversibly electrochromic when switching between theconducting (green) and the non-conducting state (blue). PANI was spincoatedonto the membrane to form the two electrodes that are separatedby the ion-conducting membrane. The MemFET and the EC display pixelwere connected with each other by silver paste. The switch from thesteady-state to dark blue occurred within 30 s, which can be consideredslow. The switching-time can be reduced, for example, by reducing theEC display pixel area (and thickness) and/or by improving the ionicconductivity and mobility. The latter argument, in particular, emphasizesthe aforementioned possibilities that are enabled by the EB technique. Forexample, by patterning the membrane and by using different electrolytesfor different devices, each device on the same membrane can beoptimized for its own purpose.Figure 3.7. (a) The schematic structure of the MemFET + EC display pixel usedin this work. (b), (c) and (d) The EC display pixel in action; (b) in the steadystate,(c) and (d) upon switching reversibly.29

Nikolai Kaihovirta3. RESULTS AND DISCUSSIONS3.3 All-Polymer Low-Voltage ECTs on Patterned Ion-Conducting MembranesPatterning of the membrane, as shown in Figure 3.5, is a simple andversatile method of controlling the ionic conductivity in the membrane. Inthis work, the technique was further elaborated and utilized fordemonstrating all-polymer ECTs manufactured on patterned ionconductingPVDF:PSSA membranes. The rather simple device structureand the low-voltage and stable operation that has been earlierdemonstrated with the PEDOT/PSS based ECTs [38, 40, 41], attracts theion-conducting membranes as suitable platforms for manufacturing ECTsby roll-to-roll manufacturing techniques.Figure 3.8. (a) The patterned PVDF:PSSA membrane. (b) Flexography is a rollto-rollsuitable printing technique. (c) A photograph of the printed ECTs on thepatterned PVDF:PSSA membrane. (d) A schematic figure of the ECT structure.The source, drain and the gate are denoted by S, D and G respectively. (e) Thetransistor channel and the gate are coupled via the PVDF:PSSA. Thus they forman electrochemical cell, as schematically shown.The membranes were prepared by the EB technique described inScheme 2.1. Besides the PVDF:PSSA, the whole ECT structure consists ofPEDOT/PSS, which is printed by flexography, as illustrated in Figure 3.8.Due to the redox-mechanism, the ECT can be fabricated in a lateralgeometry, unlike OFETs. Therefore, all three electrodes can be placed onthe same printing plate, and the device structure printed in one step.Naturally, this also reduces the issues related to electrode alignment. Inthis work, two modifications of the water-based PEDOT/PSS dispersionwere done to improve the printed structure. Firstly, the viscosity of thedispersion was adjusted by partially removing water by evaporation.Hence, the resolution of the print pattern was improved. Additionally the30

Nikolai Kaihovirta3. RESULTS AND DISCUSSIONSsolid-content of the dispersion increased, which resulted in a higherconductivity. Secondly, ethylene glycol was added to the PEDOT/PSSdispersion. It is known that the addition of a high-boiling point polarsolvent remarkably improves the conductivity in the PEDOT/PSSdispersion [67-70]. This is due to favorable morphological changes in thefilm formation that improves the charge percolation path.A schematic view of the ECT is shown in Figure 3.8 (d). The electroniccurrent between the source and drain is modulated by redox-reaction inthe transistor channel. The transistor channel and the gate are coupled viathe PVDF:PSSA and, thus, they form an electrochemical cell, as shown inFigure 3.8 (e). The reversible redox-mechanism for PEDOT/PSS isillustrated in Scheme 1.1. An increased reduction of PEDOT (PEDOT 0 ) inthe transistor channel is obtained by applying an electric potentialdifference between the gate and the transistor channel (i.e., by addressinga positive gate voltage). Consequently, the electronic current through thetransistor channel decreases. By removing the electric potential, thetransistor channel is again oxidized, as described by Scheme 1.1. In thiswork, the transistor channel area was kept between 0.9 – 1.7 mm 2 whilethe gate electrode area was kept at 50 mm 2 . Since there are more PEDOT + -than PEDOT 0 -sites in the PEDOT/PSS, the number of oxidation-sites inthe gate is increased by increasing the gate electrode area [38, 40]. Thus, ahigher on/off ratio is gained; in this work typically 10 2 – 10 3 .Three output-characteristics for a printed PEDOT/PSS based ECT aredisplayed in Figure 3.9. The characteristics are subsequently measured atdelay-times (i.e., the time-gap between each voltage step) 1 s (opensquare), 5 s (filled circle) and finally at 1 s (open triangle) again. Thelower current throughput at a higher delay-time corresponds to Equation(1.13). At longer delay-times a higher number of charges participate inreducing the transistor-channel. The minor mismatch between the firstand the last characteristic (delay-time 1 s, open symbols) has not beenclarified. It may stem from an incomplete oxidation of the transistorchannel due to slow proton diffusion. But it may also be a smalldegradation of the PEDOT caused by an oxidative degradation in thepresence of the hydrous electrolyte [71]. Oxidative degradation of P3HT(i.e., another polythiophene) in the hydrous MemFET setup will befurther studied in Sections 3.4 and 3.5 (Papers 4 and 5).The observed pinch-off of the current differs from the pinch-offmechanism for OFETs, as described in Section 1.2.2 [40]. The saturation ofthe current is caused by the reduction of PEDOT throughout the bulk ofthe transistor channel. At V G = 0 V, and at higher –V DS, protons areattracted into the transistor channel and gradually reduces the PEDOTclosest to the PVDF:PSSA-interface.31

Nikolai Kaihovirta3. RESULTS AND DISCUSSIONS50I D[A]0-50-100-150V G=1.6V0.8 V0V-1.5 -1.0 -0.5 0.0 0.5V DS[V]Figure 3.9. Output-characteristics for a PEDOT/PSS based ECT printed onto apatterned PVDF:PSSA membrane. The characteristics are subsequentlymeasured at delay-times 1 s (open square), 5 s (filled circle) and 1 s (opentriangle).PEDOT/PSS is pristinely oxidized, as shown in Scheme 1.1. Anoverpotential will irreversibly overoxidize the polymer, which results in adramatic loss of electronic conductivity [72]. Since overoxidation ofPEDOT (and other polythiophenes) causes structural changes to thepolymer and destroys the π-conjugation [71, 72], the loss in the electronicconductivity is permanent. This is shown in Figure 3.10, where a highoxidation potential (WRITE-pulse: V G = -25 V, 5 s) decreases the current(I DS) by several orders of magnitude. On the other hand, a high reductionpotential (WRITE-pulse: V G = 25 V, 5 s) does not deteriorate theconductivity in the transistor channel. The recovery seems to be almostcomplete already a few seconds after applying the WRITE-pulse. Thus,the ECTs are reversibly operated by reducing the transistor channel.32

Nikolai Kaihovirta3. RESULTS AND DISCUSSIONSV [V]20100-10-2010 -4WRITEREAD"1"abs(I D)[A]10 -610 -8"0"0 50 100 150 200t [s]Figure 3.10. The figure illustrates the difference between applying a highreduction potential versus a high oxidation potential in a PEDOT/PSS basedECT. For the former the recovery of the current is almost complete, while thelatter results in a terminated electronic conductivity. The WRITE-pulse is ±25V, 5 s and the READ-pulse is -2 V, 2 s.Overoxidation was utilized for demonstrating write-once read manytimes (WORM) memory functionality by using the ECT device structure(subsequently denoted as WORM-ECT). In a WORM-device, informationis permanently stored once and read several times. Such memory-devicesare considered favorable for e.g., archiving purposes. Organic WORMdeviceshave been demonstrated before [73-75]. I-V characteristics for theWORM-ECT are shown in Figure 3.11. By applying a high negativepotential at the gate-electrode (WRITE-pulse: -25 V, 30 s) the transistorchannel is overoxidized. In other words, the high-conducting state (“1”) istransformed into a low-conducting state (“0”). The permanent “0” state isfurther illustrated by characterizing the WORM-ECT roughly one monthafterwards. From Figure 3.11 it is apparent that the “0” state is notcompletely stable. The variation in conductivity is due to the ionicconductivity in the WORM-ECT, which is sensitive to a change in theambient humidity.33

Nikolai KaihovirtaI D[A]10 -410 -610 -810 -103. RESULTS AND DISCUSSIONS"1""0", t = 2 mins"0", t =1month-2 -1 0 1 2V DS[V]Figure 3.11. The WORM-ECT characterized before and after applying theWRITE-pulse. Minor deviations in the “0”-state occurs due to small variationsin the ambient humidity.It is noteworthy that the WRITE-period for the presented device isfairly long. This was intentionally chosen in order to ensure completeoveroxidation of the transistor channel and, hence, demonstrate theWORM memory functionality. By optimizing the volume of the transistorchannel and/or improving the ionic conductivity of the PVDF:PSSA, theWRITE-pulse could be shortened, which possibly would attract theWORM-ECT towards practical applications.3.4 The Role of Moisture in Hydrous Electrolyte GatedOFETsIn the efforts of improving the performance of the sulfonic acid-basedMemFETs (PTFE:PFSA and PVDF:PSSA), described in Section 3.2 (Paper2), it was observed that moisture management and voltage control washighly crucial in order to avoid deteriorating effects. The latter, inparticular, was somewhat surprising since the maximum electricpotential that could be applied was reached already below 1.5 V.Furthermore, one of the main arguments for using membranes for (pchannel)electrolyte gated OFETs was to avoid electrochemical interaction(See Sections 1.2.3 and 3.2) Moisture was at an early stage assumed to bethe root cause for the observed deterioration. Therefore, the effects frommoisture in hydrous electrolyte gated OFETs were clarified in this work.The work is divided in two parts: the first part deals with the reversible34

Nikolai Kaihovirta3. RESULTS AND DISCUSSIONSdeterioration and the second part presents the irreversible degradingeffects caused by the presence of moisture.3.4.1 The Effects of a Varying Ambient HumidityIn this work, PTFE:PFSA (Nafion 115) was selected as the modelmembrane, mainly because of its optical transparency. The PTFE:PFSAmembrane is commercially available and was used as received.a) b)1.0 Electrode0.5Membrane0.0Electrode0.0Electrode ElectrodeMembraneI/ A-1.0-2.0-1.5 -1.0 -0.5 0.0V/ VIn air5hrs1d5dI/ A-0.5-1.0-1.5 -1.0 -0.5 0.0Figure 3.12. Characterization at constant polarity (a) through the PTFE:PFSA,and (b) on the surface of the membrane. The measurements were done in ambientair and as a function of time in a dry atmosphere.In Figure 3.12, I-V characteristics are swept at constant polarity boththrough the PTFE:PFSA (Figure 3.12 (a)) and on the surface of themembrane (Figure 3.12 (b)). In ambient air both measurements displaysimilar behavior. At a small negative offset, there already is a strong nonlinearincrease in the current in the forward sweep (from 0.2 V to -1.2 V).The non-linear trend diminishes when exposing the samples to a dryatmosphere. The non-linear characteristic is a clear evidence for the waterelectrolysis current (i.e., Faradaic current) [76], which obviously loosessignificance upon drying the membrane. From Figure 3.12 it is also clear,that the surface of the PTFE:PFSA dries fast (Figure 3.12 (b)) compared tobulk (Figure 3.12 (a)). The I-V behavior, in Figure 3.12 (b), is completelylinear after five days of storing the sample in a dry atmosphere, intersectsat origo and has dropped by several orders of magnitude. Thus, theremaining current mainly constitutes of a small charging current andelectronic leakage. It is noteworthy that the electrical stress and theresulting water electrolysis that the samples are subjected to (in a dryV/ V35

Nikolai Kaihovirta3. RESULTS AND DISCUSSIONSatmosphere) accelerate the drying of the membrane. Moisture in themembrane is simply consumed faster when a voltage is applied over it.C i/Fcm -210 -510 -75d10 -945 mins1dIn air10 -1110 0 10 1 10 2 10 3 10 4 10 5 10 6f/ HzFigure 3.13. The C i was calculated at different frequencies from the impedancespectroscopy measurements. The measurements were done in ambient air and asa function of time in vacuum.Impedance spectroscopy was used to characterize C i as a function ofthe AC frequency, as shown in Figure 3.13. The AC amplitude was 15 mVand the frequency was scanned at zero offset bias, in order to minimizethe effect from water electrolysis. In Figure 3.13, C i is monitored in air andas a function of time in vacuum. The maximum C i , measured at lowfrequencies and in ambient air, is 5 µF/cm 2 . The capacitance-value istypically 2 – 10 times lower than usually found for electrolytes [28, 29, 31,32, 64], which corresponds to the discussion in Sections 1.2.3 and 3.2.Since the polyanion is covalently bound to the base-polymer it will bevirtually immobile. The total C i over the electrolyte will therefore belower and the electric field distribution will be broadly distributed at themembrane/P3HT interface, as schematically illustrated in Figure 1.4 (c).On the other hand, we anticipate that the immobile polyanion is hinderedfrom penetrating and electrochemically interacting with the polymericsemiconductor.In Figure 3.13, it is clear that both the characteristics and C i changeswith time in vacuum. Firstly, the C i at lower frequencies continuouslydrops with time in vacuum. Secondly, the C i becomes strongly frequencydependent in vacuum and the “knee” shifts towards lower frequencieswith longer time in vacuum. Both features correspond to the gradualdrying of the PTFE:PFSA. The hydrophilic nano-channels in thePTFE:PFSA (described in Section 3.2) shrink [62, 63], starting from thesurfaces of the membrane, which results in a poor ionic re-orientationand, hence, in a lower C i. Furthermore, the ions become unable to36

Nikolai Kaihovirta3. RESULTS AND DISCUSSIONScompletely re-orient themselves at higher frequencies. Apparently, theionic mobility drops.-I D/ A4.02.00.0a)In airb)1d5d0.0 0.5 1.0 1.52 months0.0 0.5 1.0 1.5-V D/V-V D/VFigure 3.14. (a) Typical output-characteristics measured for a PTFE:PFSA-basedMemFET in ambient air and as a function of time in a dry atmosphere. Thearrow points at the gate leakage current caused by the hydrous electrolyte. Thegate leakage current diminishes upon drying the membrane. (b) Outputcharacteristicsfor a PTFE:PFSA-based MemFET measured after two months ofstoring in a dry atmosphere. The applied V G´s are 0 V (square), -0.4 V (circle)and -0.8 V (triangle).Based on the aforementioned observations on the bare PTFE:PFSA, achanging ambient humidity should also have a detrimental impact on thecharge transport in the MemFET. Indeed, as seen in Figure 3.14, theoutput-curves loose their transistor-characteristic features (i.e., the pinchoffand the modulation) by keeping the PTFE:PFSA-based MemFET in adry atmosphere for two months. If assuming that W, L and (V G - V T ) areconstant, throughout the measurements in Figure 3.14, then the observeddecrease of the channel conductance (g D) at V DS 10 %, the PTFE:PFSA-based MemFET suffers from anincreasing ionic conductivity [76]. Therefore, careful balancing of theambient humidity levels is essential on a longer time-scale.0.30.20.10.0-I D/ A37

Nikolai Kaihovirtaa) b)3. RESULTS AND DISCUSSIONS0.04P3HTMembraneV-T/T0.00-0.04-0.08-0.12500 600 700 8001.00 V-0.50 V-0.75 V-1.00 V-1.25 V-1.50 V-2.00 V-3.00 V nmFigure 3.15. (a) A schematic illustration of the experimental setup for the opticalstudies. (b) The change in transmission was characterized for theP3HT/PTFE:PFSA sample in air at different voltages applied on the gate. The π-π* absorption region and the polaron absorption regions are indicated by thearrows.0V3.4.2 Irreversible Degradation at Higher Electric PotentialsElectronical and structural changes in the polymeric semiconductor canbe monitored in situ by registering the changes in the optical transmissionspectrum at different voltage steps [30]. In this work, the change intransmission was studied in an experimental setup as schematicallyshown in Figure 3.15 (a) (the experimental details are described in Section2.2.2). The sample resembles the MemFET-structure and the voltages areapplied on the gate electrode. The obtained spectra are plotted in Figure3.15 (b). Two regions of interest are indicated by the arrows. Firstly, the π-π* absorption region for P3HT is at 500 – 550 nm, which is the specificbandgap energy for the polymeric semiconductor. Secondly, the polaronabsorption region is at > 700 nm. Three specific regions are identified inFigure 3.15 (b). Firstly, at V G = +1 V, the P3HT is depleted from charges,which is manifested as an increased absorption (-ΔT/T > 0) in the π-π*region and a photobleaching (-ΔT/T < 0) in the polaron region. Secondly,at -1.25 V < V G < 0 V the P3HT is doped, which is seen as a smallphotobleaching in the π-π* region and an increased absorption in thepolaron region. Correspondingly, delocalized charges are induced in theP3HT. Thirdly, at V G < -1.25 V photobleaching occurs in both regions.Because of the dramatic photobleaching in the π-π* region (particularlybelow -1.5 V) there is a perturbation on the molecular level of the P3HT.38

Nikolai Kaihovirta3. RESULTS AND DISCUSSIONSa) b)5.0545 nm0.210.4 V-2.0 VV G=-0.8VT [x10 -3 ]0.200.190.18-1.6 V0.4 V-1.2 V-0.8 V -1.0 V0V -0.4 V0.4 V0.4 V 0.4 V0.4 V 0.4 V00:00 01:00 02:00t/ hrs:min-I D/ A2.50.0-1.2 V-1.6 V0.0 0.5 1.0 1.5-V D/V-0.4 V0VFigure 3.16. (a) The optical transmission recorded in the π-π* absorption regionfor the P3HT/PTFE:PFSA sample in air. The transmission is monitored atdifferent gate voltage steps. The experimental setup is shown in Figure 3.15 (a).(b) Output-characteristics for a PTFE:PFSA-based MemFET in air.In Figure 3.16 (a) the transmission (T) in the π-π* region is studied byapplying a sequence of voltage steps. The system is dedoped by a positivebias of 0.4 V, according to the off-state of the PTFE:PFSA-based MemFET(Figure 3.3 (c)). The optical behavior is further compared to the typicaloutput-characteristic measured for a PTFE:PFSA-based MemFET (Figure3.16 (b)). Both measurements are done in air. However, the MemFETcharacterization is performed in a slightly drier air (RH = 9 – 10 %),compared to the optical characterizations (RH = 25 – 30 %), in order todiminish the influence from gate leakage currents. The observed trend inFigure 3.15 is also seen in Figure 3.16. At -1.2 V < V G < 0 V, the change inT is small, relatively fast and reversible. Furthermore, T remains constantupon applying the doping-bias and the change is nearly equal for all thebias steps within the interval. All these facts correspond to theelectrostatic doping of the P3HT, which is verified by the OFETcharacteristiccurves (Figure 3.16 (b)) within the same interval. On theother hand, at V G < -1.2 V, the change in T becomes large, continuouslygrowing and irreversible. In the output-characteristic this is identified asa clear deterioration of the modulation and the pinch-off. The system hasevidently entered a regime of irreversible degradation.The observed degradation in Figures 3.15 and 3.16 can be furtherdistinguished by studying a sequence of several transfer-sweepscharacterized in a dry atmosphere. Figure 3.17 displays 25 consecutivelymeasured transfer-sweeps for PTFE:PFSA-based MemFETs. In Figure 3.17(a) the transfer-measurements are carried out in the saturated region,while they are characterized in the linear region in Figure 3.17 (b).39

Nikolai Kaihovirtaa) b)3. RESULTS AND DISCUSSIONS10-5V D =-1.5V1st sweep1st sweep25th10 -610 -6I D/ A10 -725thV D=-0.4V10 -7I D/ A10 -8-2 -1 0 1V G/V-1.5 -1.0 -0.5 0.0 0.5V G/VFigure 3.17. 25 consecutive forward scans for PTFE:PFSA-based MemFETsmeasured (a) in the saturated region, and (b) in the linear region.Moreover, in Figure 3.17 (b), the electric potential never exceeds 1.2 Vbetween any two electrodes. Based on the results it can be concluded thatthe onset for the degradation of the transfer-curve, in Figure 3.17 (a), isbias-dependent. These results correspond with the aforementionedobservations in Figures 3.15 and 3.16.SC 6 H 13-e - +H 2O+e - S +S+ e - + 2H + DiketonesSulfonesC 6 H 13C 6 H 13OCarboxylic acidsScheme 3.1. The obtained results in Section 3.4.2 support the following reactionmechanism. Reversible doping of the P3HT occurs at -1.2 V < V G < 0 V for thePTFE:PFSA-based MemFET. At V G < -1.2 V, oxidative degradation of theP3HT will be initiated by significant water electrolysis in the hydrouselectrolyte. The sulfoxide group will further react and the semiconductingproperties of the P3HT will be lost.The experimental results obtained in this work point towards thedegradation of polythiophenes by oxidation, which is initiated by waterelectrolysis in the hydrous electrolyte (Scheme 3.1). For the PTFE:PFSAbasedMemFET the onset for severe oxidative degradation is at V G ≈ -1.2V. (Over)oxidative degradation of polythiophene-derivatives has beenstudied before [71, 72, 77-79]. For example, according to Barsch, et. al.[71], the complete degradation mechanism consists of several steps that40

Nikolai Kaihovirta3. RESULTS AND DISCUSSIONSwill result in chain breaking and in the loss of conjugation. In otherwords, the conducting pathway will be disrupted.3.5 Improved Device Stability by Adding a StericallyHindered Phenolic Antioxidant in P3HTIn Section 3.4 (Paper 4) it was shown that the P3HT is susceptible todegradation by oxidation at electric overpotentials in a hydrouselectrolyte gated OFET. This will, of course, have detrimentalconsequences for the long-term stability of the transistor. There are threealternatives for reducing the degradation: (1) by choosing a lesshygroscopic electrolyte. For example, the HIFET can be operated at abroader electric potential interval and at higher RH levels than thePTFE:PFSA-based MemFET. (2) By controlling the electric potentialinterval. This was shown in Figure 3.17. (3) By protecting the polymericsemiconductor from oxidative degradation with a sterically hinderedphenolic antioxidant. This is a well known method for inhibitingpolymeric decomposition of conventional polyolefins [80]. In this sectionit will be demonstrated that sterically hindered phenolic antioxidants canalso be used for improving the stability of π-conjugated polymers.a) b)C 6 H 13SHOnS C 6 H 13OHc) d) P3HT (+ stabilizer)CF 2CF 2mCF 2CFOCFCF 2pOnCF 3OCF 2CF 2S OOHPTFE:PFSAFigure 3.18. The chemical structures of (a) the P3HT-semiconductor, (b) the4,4´-methylenebis(2,6-di-tert-butyl phenol)-stabilizer, and (c) the PTFE:PFSAmembrane.(d) The structure for the MemFET.SGD41

Nikolai Kaihovirta3. RESULTS AND DISCUSSIONSThe PTFE:PFSA-based MemFET is used as the model device (Figure3.18). The sterically hindered phenolic antioxidant 4,4´-methylenebis(2,6-di-tert-butyl phenol) (subsequently denoted as stabilizer) is added intothe P3HT-solution. The chemical structure for the stabilizer is shown inFigure 3.18 (b) and the stabilization mechanism is illustrated in Scheme3.2. Accordingly, the stabilizer structure serves several purposes [80].Highly reactive alkoxy and peroxy radicals are stabilized by the easilyabstractable hydrogen from the phenolic moiety. Furthermore, theantioxidant complex is stabilized by the resonance structure, provided bythe aromatic system. The alkyl substituents at 2,6-positions provide twofunctionalities. Firstly, they enhance the hydrogen donating ability of theantioxidant. Secondly, they hinder the formed phenoxyl radical to furtherreact by abstracting hydrogen from the polymeric backbone.COOHROOOCO+ROOHOScheme 3.2. The stabilization mechanism for sterically hindered phenolicantioxidants.P3HT- and P3HT+stabilizer-samples were fabricated by spin-coatingthe solutions on respective glass substrates. Thermal stress studies werecarried out in ambient air and in darkness at 110 ºC. The change in theUV-Vis absorption was regularly monitored on the samples in order toobserve any signs for improved lifetime in the stabilized P3HT. Theresults are depicted in Figure 3.19. The change in absorbance is similar inboth cases. Initially, the annealing of the samples will improve theabsorption but after a few hours of annealing both samples displaybleaching of the absorbance peak. Further thermal stress will result inblue-shift of the peak and eventually to a complete loss of the absorbance[81]. Even though the bleaching seemingly falls with equal rates, afterfour hours of annealing, the absorbance is consistently higher throughoutC42

Nikolai Kaihovirta3. RESULTS AND DISCUSSIONSthe annealing process. Apparently, the stabilization occurs within the firstfew hours of annealing, which is verified by Figure 3.19 (b). Two featuresare distinguished in the stabilized P3HT. Firstly, the change inabsorbance is relatively higher and, secondly, the improved absorbanceoccurs within two hours of annealing whereas the absorbance for theunstabilized P3HT decreases already after 30 minutes of annealing.a)3.0Absorbanceb)Absorbance [norm.]2.01.00.01.21.11.00.9Dash: P3HTSolid: P3HT + stabilizer400 600 800 = 518 nmWavelength [nm]0hrs1hrs4hrs40 hrs1.000 1 2Time of exposure [hrs]P3HTP3HT + stabilizer0 10 20 30 40Time of exposure [hrs]Figure 3.19. (a) UV-Vis absorption spectra of stabilized and unstabilized P3HTsamplesupon annealing at 110 ºC in air. (b) The normalized change inabsorbance at 518 nm and as a function of time of annealing. The inset shows thechange in absorbance for the first two hours.The results in Figure 3.19 support that two major effects contribute tothe changes in the absorbance: (1) densification of the P3HT-lattices [51,82], and (2) thermooxidation of the P3HT [81]. The former characterizesthe initial rise in the absorbance while the latter dominates at long-termannealing in air. Hence, in the stabilized P3HT, the polymer is initiallyprotected from thermooxidation by the sterically hindered phenolicantioxidant. Eventually, however, the stabilizer will be consumed and the1.151.101.05Absorbance [norm.]43

Nikolai Kaihovirta3. RESULTS AND DISCUSSIONSthermooxidation will continue with the same rate compared to theunstabilized P3HT.10 -61st1st1st10 -7 100th100th10 -8V d=-1.5V 100th-3 -2 -1 0 1 -3 -2 -1 0 1 -3 -2 -1 0 110 -5 V g[V]V 0V g[V]V g[V]a)b)c)10 -5-I d[A]Figure 3.20. 100 consecutive transfer-characteristics measured in the saturatedregion for (a) an unstabilized MemFET and for MemFETs having a stabilizerconcentration of (b) 500 ppm by weight and (c) 1:1 by weight. Only the forwardsweeps are shown. The first (green) and the 100th sweep (red) are also indicated.The used stabilizer concentration in the UV-Vis analysis is 1:1 byweight (relative to the P3HT). From the observations in Figure 3.19 it isanticipated that even higher concentrations would further improve thelifetime of the P3HT. However, such concentrations of an “impurity” willnaturally affect the charge transport in the P3HT in an unfavorablefashion. In Figure 3.20, 100 consecutive transfer-sweeps, measured in thesaturated region, are shown for an unstabilized MemFET (3.20 (a)) andfor MemFETs having a stabilizer concentration of 500 ppm by weight(3.20 (b)) and 1:1 by weight (3.20 (c)). Clearly, adding the stabilizerdecreases the pristine on-current. Furthermore, by calculating µ,according to Equation (1.10) (assuming that W, L and C i are identical forthe devices), the difference between the first sweeps in Figures 3.20 (a)and (c) is two orders of magnitude. Nevertheless, the fall of the on-statethroughout the 100 sweeps is at its minimum in Figure 3.20 (c). In otherwords, the stabilizer protects the P3HT from oxidative degradation thatwould be caused by the applied electric overpotentials, as described inSection 3.4 (Paper 4). Furthermore, Figure 3.20 (b) demonstrates that thenegative effects on the electronic performance can be reduced bylowering the stabilizer concentration without loosing the stabilizingeffect.In Scheme 3.2 it was shown that oxidative degradation of the P3HTresults in a disrupted charge percolation. This would imply that a higherelectric potential is required in order to accumulate mobile charges andestablish the active channel between the source and drain. A highernegative shift of V 0 is indeed distinguished for the unstabilized MemFET(see Figure 3.20). Correspondingly, when estimating V T from Equation10 -610 -710 -8-I d[A]44

Nikolai Kaihovirta3. RESULTS AND DISCUSSIONS(1.4) a similar trend was also observed. However, the analysis providedsomewhat contradictory results, which is ascribed to the decaying C ithroughout the 100 transfer-sweeps. Water electrolysis occurs at highervoltages and since the devices are measured in a dry atmosphere themoisture will eventually be depleted. With prolonged scans the ionicpolarization in the membrane is deteriorated, which, according toEquation (1.4), affects the slope and the on-current in the transfercharacteristics.It should be emphasized that this effect should beidentical for all MemFETs in Figure 3.20. Noteworthy is that, for thestabilized MemFET in Figure 3.20 (c), V 0 shifts towards positive voltagesafter roughly 30 transfer-sweeps. The effect is probably caused by theincreasing concentration of phenoxyl radicals, in the P3HT, acting asunintentional dopants. The effect is not visible in Figure 3.20 (b).1.0I d[norm.]0.5P3HTP3HT + stabilizer0.00 6 12Time [hrs]Figure 3.21. The normalized I DS as a function of time with constant applied V DS= -1.5 V and V G = -1.0 V.The problems associated to a decaying C i can be avoided by applying aconstant V DS and V G . Thus, a high charge density is continuouslyconcentrated in the active channel. In Figure 3.21, the I-t characteristicsare monitored in the on-state for an unstabilized and a stabilizedMemFET. Accordingly, the on-current decreases more rapidly in theunstabilized device, compared to the stabilized MemFET, whichcorresponds to the oxidative degradation of P3HT. Such degradation isnot observed for the stabilized MemFET and the obtained results inFigure 3.21 are yet another evidence for the protective effect of thesterically hindered phenolic antioxidants against P3HT degradation.45

Nikolai Kaihovirta4. SUMMARY4. SUMMARYThe reported results in this thesis have broadened our knowledgeconcerning ion modulated organic transistors and resulted in new ideastowards printed organic electronics. The previous work on the HIFETconcept,conducted in our laboratory, was used as the starting-point. Bydirectly comparing ordinary OFETs with HIFETs we have been able toshow that the HIFET (as well as other electrolyte gated OFETs) are lesssensitive to an increased roughness of the substrate. Consequently, allprintedand low-voltage HIFETs have been demonstrated on both plasticand paper substrates.We have also demonstrated ion modulated organic transistors built onion-conducting membranes (solid electrolytes). The membrane functionsas a platform for fabricating the devices. Moreover, by using the EBtechnique, the membranes can be functionalized and tailor-made in anydesired fashion.The role of water in hydrous electrolytes has been clarified. Firstly,unlike traditional OFETs, the capacitance should often be considered as avariable because of the varying humidity level in the hydrous electrolyte“insulator”. Secondly, moisture has both a reversible and an irreversibledeteriorating effect on the device behavior. The former can be largelyhandled by controlling the ambient humidity. The latter degradation, onthe other hand, can be diminished by using a less hygroscopic electrolyteand/or by carefully controlling the electric potentials in the device.Alternatively, we have shown that the degradation can be quenched byadding a sterically hindered phenolic antioxidant into the polymericsemiconductor.In future research, improved consistency and device stability shouldbe emphasized. These key-issues could be more essential for printedapplications than e.g., the efforts towards “state-of-the-art” chargemobilities or on/off ratios. In this thesis, the ion-conducting membraneshave been mainly associated with ion modulated organic transistors. Infuture, the concept will be further broadened. For example, varioussensing applications might enable new possibilities for research. The ionmodulated organic transistor could be built on the same membrane andutilized e.g., as a signal amplifier. Thus, the multifunctional properties ofthe membranes and the ion-to-electron transducing opportunities wouldbe brought up to new levels.47

Nikolai KaihovirtaBIBLIOGRAPHYBIBLIOGRAPHY[1] J. Bardeen, W. H. Brattain. The Transistor, a Semi-ConductorTriode. Phys Rev. 74, 230-231 (1948).[2] D. Kahng. Electric Field Controlled Semiconductor Device. USPatent 3 102 230 (1963).[3] C. K. Chiang, C. R. Fincher, Y. W. Park, A. J. Heeger, H. Shirakawa,E. J. Louis, S. C. Gau, A. G. MacDiarmid. Electrical Conductivity inDoped Polyacetylene. Phys. Rev. Lett. 39, 1098-1101 (1977).[4] H. S. White, G. P. Kittlesen, M. S. Wrighton. ChemicalDerivatization of an Array of Three Gold Microelectrodes withPolypyrrole: Fabrication of a Molecule-Based Transistor. J. Am.Chem. Soc. 106, 5375-5377 (1984).[5] A. Tsumura, H. Koezuka, T. Ando. Macromolecular electronicdevice: Field-effect transistor with a polythiophene thin film. Appl.Phys. Lett. 49, 1210-1212 (1986).[6] M. Berggren, A. Richter-Dahlfors. Organic Bioelectronics. Adv.Mater. 19, 3201-3213 (2007).[7] R. A. Street, A. Salleo. Contact effects in polymer transistors. Appl.Phys. Lett. 81, 2887-2889 (2002).[8] Y. Roichman, N. Tessler. Structures of polymer field-effecttransistor: Experimental and numerical analyses. Appl. Phys. Lett. 80,151-153 (2002).[9] T. J. Richards, H. Sirringhaus. Analysis of the contact resistance instaggered, top-gate organic field-effect transistors. J. Appl. Phys. 102,094510 (2007).[10] L.-L. Chua, P. K. H. Ho, H. Sirringhaus, R. H. Friend. High-stabilityultrathin spin-on benzocyclobutene gate dielectric for polymer fieldeffecttransistors. Appl. Phys. Lett. 84, 3400-3402 (2004).49

Nikolai KaihovirtaBIBLIOGRAPHY[11] Y.-Y. Noh, N. Zhao, M. Caironi, H. Sirringhaus. Downscaling ofself-aligned, all-printed polymer thin-film transistors. Nat.Nanotechnol. 2, 784-789 (2007).[12] L. Burgi, R. Pfeiffer, C. Winnewisser. Submicrometer polymertransistors fabricated by a mask-free photolithographic selfalignmentprocess. Appl. Phys. Lett. 92, 153302 (2008).[13] J. N. Haddock, X. Zhang, S. Zheng, Q. Zhang, S. R. Marder, B.Kippelen. A comprehensive study of short channel effects in organicfield-effect transistors. Org. Electron. 7, 45-54 (2006).[14] C. Tanase, E. J. Meijer, P. W. M. Blom, D. M. de Leeuw. Unificationof the Hole Transport in Polymeric Field-Effect Transistors andLight-Emitting Diodes. Phys. Rev. Lett. 91, 216601 (2003).[15] C. D. Dimitrakopoulos, S. Purushothaman, J. Kymissis, A. Callegari,J. M. Shaw. Low-Voltage Organic Transistors on Plastic ComprisingHigh-Dielectric Constant Gate Insulators. Science 283, 822-824(1999).[16] C. Bartic, H. Jansen, A. Campitelli, S. Borghs. Ta 2O 5 as gate dielectricmaterial for low-voltage organic thin-film transistors. Org. Electron.3, 65-72 (2002).[17] G. Wang, D. Moses, A. J. Heeger, H.-M. Zhang, M. Narasimhan, R.E. Demaray. Poly(3-hexylthiophene) field-effect transistors withhigh dielectric constant gate insulator. J. Appl. Phys. 95, 316- (2004).[18] M. Halik, H. Klauk, U. Zschieschang, G. Schmid, W. Radlik, W.Weber. Polymer Gate Dielectrics and Conducting-Polymer Contactsfor High-Performance Organic Thin-Film Transistors. Adv. Mater.14, 1717-1722 (2002).[19] M. E. Roberts, S. C. B. Mannsfeld, N. Queralto, C. Reese, J. Locklin,W. Knoll, Z. Bao. Water-stable organic transistors and theirapplication in chemical and biological sensors. P. Natl. Acad. Sci.105, 12134-12139 (2008).50

Nikolai KaihovirtaBIBLIOGRAPHY[20] Y.-Y. Noh, H. Sirringhaus. Ultra-thin polymer gate dielectrics fortop-gate polymer field-effect transistors. Org. Electron. 10, 174-180(2009).[21] J. Collet, O. Tharaud, A. Chapoton, D. Vuillaume. Low-voltage, 30nm channel length, organic transistors with a self-assembledmonolayer as gate insulating films. Appl. Phys. Lett. 76, 1941-1943(2000).[22] L. A. Majewski, R. Schroeder, M. Grell. Low-Voltage, High-Performance Organic Field-Effect Transistors with an Ultra-ThinTiO 2 Layer as Gate Insulator. Adv. Funct. Mater. 15, 1017-1022 (2005).[23] H. Klauk, U. Zschieschang, J. Pflaum, M. Halik. Ultralow-powerorganic complementary circuits. Nature, 445, 745-748 (2007).[24] J. Bobacka, A. Ivaska, M. Grzeszczuk. Electrochemical study ofpoly(3-octylthiophene) film electrodes II. Reversibleredox/conductivity state switching. Impedance study. Synth. Met.44, 21-34 (1991).[25] M. J. Panzer, C. D. Frisbie. High Carrier Density and MetallicConductivity in Poly(3-hexylthiophene) Achieved by ElectrostaticCharge Injection. Adv. Funct. Mater. 16, 1051-1056 (2006).[26] H. G. O. Sandberg, T. G. Bäcklund, R. Österbacka, H. Stubb. High-Performance All-Polymer Transistor Utilizing a HygroscopicInsulator. Adv. Mater. 16, 1112-1115 (2004).[27] T. G. Bäcklund, R. Österbacka, H. Stubb, J. Bobacka, A. Ivaska.Operating principle of polymer insulator organic thin-filmtransistors exposed to moisture. J. Appl. Phys. 98, 074504 (2005).[28] L. Herlogsson, X. Crispin, N. D. Robinson, M. Sandberg, O.-J. Hagel,G. Gustafsson, M. Berggren. Low-Voltage Polymer Field-EffectTransistors Gated via a Proton Conductor. Adv. Mater. 19, 97-101(2007).[29] J. H. Cho, J. Lee, Y. Xia, B. Kim, Y. He, M. J. Renn, T. P. Lodge, C. D.Frisbie. Printable ion-gel gate dielectrics for low-voltage polymerthin-film transistors on plastic. Nat. Mater. 7, 900-906 (2008).51

Nikolai KaihovirtaBIBLIOGRAPHY[30] J. D. Yuen, A. S. Dhoot, E. B. Namdas, N. E. Coates, M. Heeney, I.McCulloch, D. Moses, A. J. Heeger. Electrochemical Doping inElectrolyte-Gated Polymer Transistors. J. Am. Chem. Soc. 129, 14367-14371 (2007).[31] J. Lee, L. G. Kaake, J. H. Cho, X.-Y. Zhu, T. P. Lodge, C. D. Frisbie.Ion Gel-Gated Polymer Thin-Film Transistors: OperatingMechanism and Characterization of Gate Dielectric Capacitance,Switching Speed, and Stability. J. Phys. Chem. C 113, 8972-8981(2009).[32] E. Said, X. Crispin, L. Herlogsson, S. Elhag, N. D. Robinson, M.Berggren. Polymer field-effect transistor gated via apoly(styrenesulfonic acid) thin film. Appl. Phys. Lett. 89, 143507(2006).[33] E. W. Paul, A. J. Ricco, M. S. Wrighton. Resistance of PolyanilineFilms as a Function of Electrochemical Potential and the Fabricationof Polyaniline-Base Microelectronic Devices. J. Phys. Chem. 89, 1441-1447 (1985).[34] J. W. Thackeray, H. S. White, M. S. Wrighton. Poly(3-methylthiophene)-Coated Electrodes: Optical and ElectricalProperties as a Function of Redox Potential and Amplification ofElectrical and Chemical Signals Using Poly(3-methylthiophene)-Based Microelectrochemical Transistors. J. Phys. Chem. 89, 5133-5140(1985).[35] S. Chao, M. S. Wrighton. Solid-State Microelectrochemistry:Electrical Characteristics of a Solid-State MicroelectrochemicalTransistor Based on Poly(3-methylthiophene). J. Am. Chem. Soc. 109,2197-2199 (1987).[36] S. Chao, M. S. Wrighton. Characterization of a ”Solid-State”Polyaniline-Base Transistor: Water Vapor Dependent Characteristicsof a Device Employing a Poly(vinyl alcohol)/Phosphoric AcidSolid-State Electrolyte. J. Am. Chem. Soc. 109, 6627-6631 (1987).[37] D. Ofer, R. M. Crooks, M. S. Wrighton. Potential Dependence of theConductivity of Highly Oxidized Polythiophenes, Polypyrroles, andPolyaniline: Finite Windows of High Conductivity. J. Am. Chem. Soc.112, 7869-7879 (1990).52

Nikolai KaihovirtaBIBLIOGRAPHY[38] D. Nilsson, N. Robinson, M. Berggren, R. Forchheimer.Electrochemical Logic Circuits. Adv. Mater. 17, 353-358 (2005).[39] A. Kumar, J. Sinha. Electrochemical Transistors for Applications inChemical and Biological Sensing. In: D. A. Bernards, R. M. Owens,G. G. Malliaras (Eds.), Organic Semiconductors in SensorApplications. Springer-Verlag. Berlin, Heidelberg, New York (2008)pp. 245-261.[40] M. Berggren, R. Forchheimer, J. Bobacka, P.-O. Svensson, D.Nilsson, O. Larsson, A. Ivaska. PEDOT:PSS-Based ElectrochemicalTransistors for Ion-to-Electron Transduction and Sensor SignalAmplification. In: D. A. Bernards, R. M. Owens, G. G. Malliaras(Eds.), Organic Semiconductors in Sensor Applications. Springer-Verlag. Berlin, Heidelberg, New York (2008) pp. 263-280.[41] M. Nikolou, G. G. Malliaras. Applications of Poly(3,4-Ethylenedioxythiophene) Doped With Poly(Styrene Sulfonic Acid)Transistors in Chemical and Biological Sensors. Chem. Rec. 8, 13-22(2008).[42] R. Bollström, A. Määttänen, D. Tobjörk, P. Ihalainen, N. Kaihovirta,R. Österbacka, M. Toivakka, J. Peltonen. A multilayer coated fiberbasedsubstrate suitable for printed functionality. Org. Electron. 10,1020-1023 (2009).[43] H. Sirringhaus, N. Tessler, D. S. Thomas, P. J. Brown, R. H. Friend,In Advances in Solid State Physics Ed. B. Kramer, Friedr. Vieweg &Sohn: Braunschweig, Vol. 39, pp. 101-110 (1999).[44] A. Salleo. Charge transport in polymeric transistors. Mater. Today 10,38-45 (2007).[45] I. Osaka, R. D. McCullough. Advances in Molecular Design andSynthesis of Regioregular Polythiophenes. Acc. Chem. Res. 41, 1202-1214 (2008).[46] R. D. McCullough, R. D. Lowe. Enhanced electrical conductivity inregioselectively synthesized poly(3-alkylthiophenes). J. Chem. Soc.,Chem. Commun. 70-72 (1992).53

Nikolai KaihovirtaBIBLIOGRAPHY[47] H. Sirringhaus, P. J. Brown, R. H. Friend, M. M. Nielsen, K.Bechgaard, B. M. W. Langevald-Voss, A. J. H. Spiering, R. A. J.Janssen, E. W. Meijer, P. Herwig, D. M. de Leeuw. Two-dimensionalcharge transport in self-organized, high-mobility conjugatedpolymers. Nature 401, 685-688 (1999).[48] R. Österbacka, C. P. An, X. M. Jiang, Z. V. Vardeny. Two-Dimensional Electronic Excitations in Self-Assembled ConjugatedPolymer Nanocrystals. Science 287, 839-842 (2000).[49] H. Sirringhaus, N. Tessler, D. S. Thomas, R. H. Friend. IntegratedOptoelectronic Devices Based on Conjugated Polymers. Science 280,1741-1744 (1998).[50] J.-F. Chang, B. Sun, D. W. Breiby, M. M. Nielsen, T. I. Sölling, M.Giles, I. McCulloch, H. Sirringhaus. Enhanced Mobility of Poly(3-hexylthiophene) Transistors by Spin-Coating from High-Boiling-Point Solvents. Chem. Mater. 16, 4772-4776 (2004).[51] M. Surin, P. Leclére, R. Lazzaroni, J. D. Yuen, G. Wang, D. Moses, A.J. Heeger, S. Cho, K. Lee. Relationship between the microscopicmorphology and the charge transport properties in poly(3-hexylthiophene) field-effect transistors. J. Appl. Phys. 100, 033712(2006).[52] T. Lehtinen, G. Sundholm, S. Holmberg, F. Sundholm, P. Björnbom,M. Bursell. Electrochemical characterization of PVDF-based protonconducting membranes for fuel cells. Electrochim. Acta 43, 1881-1890(1998).[53] N. Walsby, M. Paronen, J. Juhanoja, F. Sundholm. Sulfonation ofStyrene-Grafted Poly(vinylidene fluoride) Films. J. Appl. Polym. Sci.81, 1572-1580 (2001).[54] N. J. Kaihovirta, D. Tobjörk, T. Mäkelä, R. Österbacka. Low-VoltageOrganic Transistors Fabricated Using Reverse Gravure Coating onPrepatterned Substrates. Adv. Eng. Mater. 10, 640-643 (2008).[55] L.-L. Chua, P. K. H. Ho, H. Sirringhaus, R. H. Friend. Observation ofField-Effect Transistor Behavior at Self-Organized Interfaces. Adv.Mater. 16, 1609-1615 (2004).54

Nikolai KaihovirtaBIBLIOGRAPHY[56] F. Xue, Z. Liu, Y. Su, K. Varahramyan. Inkjet printed silversource/drain electrodes for low-cost polymer thin film transistors.Microel. Eng. 83, 298-302 (2006).[57] F. A. Yildrim, R. R. Schliewe, W. Bauhofer, R. M. Meixner, H.Goebel, W. Krautschneider. Gate insulators and interface effects inorganic thin-film transistors. Org. Electron. 9, 70-76 (2008).[58] Y. Jung, R. J. Kline, D. A. Fischer, E. K. Lin, M. Heeney, I.McCulloch, D. M. DeLongchamp. The Effect of InterfacialRoughness on the Thin Film Morphology and Charge Transport ofHigh-Performance Polythiophenes. Adv. Funct. Mater. 18, 742-750(2008).[59] D. Tobjörk, N. J. Kaihovirta, T. Mäkelä, F. S. Pettersson, R.Österbacka. All-printed low-voltage organic transistors. Org.Electron. 9, 931-935 (2008).[60] Y. Xia, J. Cho, B. Paulsen, C. D. Frisbie, M. J. Renn. Correlation ofon-state conductance with referenced electrochemical potential inion gel gated polymer transistors. Appl. Phys. Lett. 94, 013304 (2009).[61] M. Hamedi, L. Herlogsson, X. Crispin, R. Marcilla, M. Berggren, O.Inganäs. Fiber-Embedded Electrolyte-Gated Field-Effect Transistorsfor e-Textiles. Adv. Mater. 21, 573-577 (2009).[62] K. A. Mauritz, R. B. Moore. State of Understanding of Nafion. Chem.Rev. 104, 4535-4585 (2004).[63] K. Schmidt-Rohr, Q. Chen. Parallell cylindrical water nanochannelsin Nafion fuel-cell membranes. Nat. Mater. 7, 75-83 (2008).[64] O. Larsson, E. Said, M. Berggren, X. Crispin. Insulator PolarizationMechanisms in Polyelectrolyte-Gated Organic Field-EffectTransistors. Adv. Funct. Mater. 19, 3334-3341 (2009).[65] M. J. Panzer, C. D. Frisbie. Polymer Electrolyte-Gated Organic Field-Effect Transistors: Low-Voltage, High-Current Switches for OrganicElectronics and Testbeds for Probing Electrical Transport at HighCharge Carrier Density. J. Am. Chem. Soc. 129, 6599-6607 (2007).55

Nikolai KaihovirtaBIBLIOGRAPHY[66] P. M. Beaujuge, J. R. Reynolds. Color Control in π-ConjugatedOrganic Polymers for Use in Electrochromic Devices. Chem. Rev.110, 268-320 (2010).[67] S. K. M. Jönsson, J. Birgerson, X. Crispin, G. Greczynski, W.Osikowicz, A. W. Denier van der Gon, W. R. Salaneck, M. Fahlman.The effects of solvents on the morphology and sheet resistance inpoly(3,4-ethylenedioxythiophene)-polystyrenesulfonic acid(PEDOT-PSS) films. Synth. Met. 139, 1-10 (2003).[68] J. Ouyang, Q. Xu, C.-W. Chu, Y. Yang, G. Li, J. Shinar. On themechanism of conductivity enhancement in poly(3,4-ethylenedioxythiophene):poly(styrene sulfonate) film throughsolvent treatment. Polymer 45, 8443-8450 (2004).[69] X. Crispin, F. L. E. Jakobsson, A. Crispin, P. C. M. Grim, P.Andersson, A. Volodin, C. van Haesendonck, M. Van derAuweraer, W. R. Salaneck, M. Berggren. The Origin of the HighConductivity of Poly(3,4-ethylenedioxythiophene)-Poly(styrenesulfonate) (PEDOT-PSS) Plastic Electrodes. Chem.Mater. 18, 4354-4360 (2006).[70] A. M. Nardes, M. Kemerink, M. M. de Kok, E. Vinken, K. Maturova,R. A. J. Janssen. Conductivity, work function, and environmentalstability of PEDOT:PSS thin films treated with sorbitol. Org.Electron. 9, 727-734 (2008).[71] U. Barsch, F. Beck. Anodic overoxidation of polythiophenes in wetacetonitrile electrolytes. Electrochim. Acta 41, 1761-1771 (1996).[72] M. Lapkowski, A. Pron. Electrochemical oxidation of poly(3,4-ethylenedioxythiophene) – ”in situ” conductivity and spectroscopicinvestigations. Synth. Met. 110, 79-83 (2000).[73] S. Möller, C. Perlov, W. Jackson, C. Taussig, S. R. Forrest. Apolymer/semiconductor write-once read-many-times memory.Nature 426, 166-169 (2003).[74] E. Y. H. Teo, Q. D. Ling, Y. Song, Y. P. Tan, W. Wang, E. T. Kang, D.S. H. Chan, C. Zhu. Non-volatile WORM memory device based onan acrylate polymer with electron donating carbazole pendantgroups. Org. Electron. 7, 173-180 (2006).56

Nikolai KaihovirtaBIBLIOGRAPHY[75] J. Lin, D. Ma. The morphology control of pentacene for write-onceread-many-timesmemory devices. J. Appl. Phys. 103, 024507 (2008).[76] E. Said, O. Larsson, M. Berggren, X. Crispin. Effects of the IonicCurrents in Electrolyte-gated Organic Field-Effect Transistors. Adv.Funct. Mater. 18, 3529-3536 (2008).[77] H. Harada, T. Fuchigami, T. Nonaka. Degradation and itsprevention, and the deactivation and reactivation of electroactivepolythiophene films during oxidation/reduction cycles. J.Electroanal. Chem. 303, 139-150 (1991).[78] N. Ljungqvist, T. Hjertberg. Oxidative Degradation of Poly(3-octylthiophene). Macromolecules 28, 5993-5999 (1995).[79] P. Tehrani, A. Kanciurzewska, X. Crispin, N. D. Robinson, M.Fahlman, M. Berggren. The effect of pH on the electrochemical overoxidationin PEDOT:PSS films. Solid State Ionics 177, 3521-3527(2006).[80] R. F. Grossman. Antioxidants. In: J. T. Lutz, Jr., R. F. Grossman(Eds.), Polymer Modifiers and Additives. Marcel Dekker, Inc. NewYork, NY (2001) pp. 1-34.[81] M. Manceau, A. Rivaton, J.-L. Gardette, S. Guillerez, N. Lemaitre.The mechanism of photo- and thermooxidation of poly(3-hexylthiophene) (P3HT) reconsidered. Polym. Degrad. Stabil. 94, 898-907 (2009).[82] B. A. Mattis, P. C. Chang, V. Subramanian. Performance recoveryand optimization of poly(3-hexylthiophene) transistors by thermalcycling. Synt. Met. 156, 1241-1248 (2006).57

Nikolai KaihovirtaSVENSK RESUMÉSVENSK RESUMÉI likhet med konventionella polymerer (plaster) är de π-konjugeradepolymererna flexibla, lösliga och processbara vid låga temperaturer (

Nikolai KaihovirtaSVENSK RESUMÉOFETer är det känt att övergången till grova substrat (såsom plast ochpapper) har försämrande effekter på transistorfunktionen. Grova substratär dock oundvikliga ifall man ämnar tillverka flexibel elektronik på ettkostnadseffektivt sätt. I avhandlingen har HIFETen studerats för attklargöra att jonmodulerade OFETer påverkas mindre utav övergångentill grövre plastsubstrat jämfört med OFETer. Med den kunskapen har visedermera presenterat lågspännings HIFETar tillverkade medmasstillverkningsmetoder på såväl plast- som på papperssubstrat.Lejonparten av avhandlingen fokuseras på jonmodulerade organiskatransistorer som utnyttjar jonledande polymera membran. Membranenhar vanligtvis tillämpats i bränslecells-, gas och vätskeseparerings- samt ikloralkali-industrin. I avhandlingen visas dock att membranet ävenfungerar som substrat och som elektrolyt för en eller flera komponenter.MemFETen (Membrane based Field-Effect Transistor) utgörmodelltransistorn men även den helpolymera och membran baseradeECTn demonstreras. I avhandlingens senare del studeras fuktens roll ielektrolyten och hur den påverkar transistorbeteendet. Samtligaelektrolyter i avhandlingen (inklusive poly(vinyl fenolen)) behöver fuktför att kunna leda joner. MemFETen används i studierna och fuktensreversibla och irreversibla inverkan på den polymera halvledarenklargörs även med hjälp av spektroskopiska tekniker. För att förhindraden irreversibla degraderingen av den polymera halvledaren presenterassteriskt hindrade fenoliska antioxidanter som ett intressant alternativ.Avhandlingen klargör skillnaderna mellan OFETen, jonmoduleradeOFETen och ECTn. De jonmodulerade organiska transistorernasmöjlighet att framställas med masstillverkningsmetoder presenteras. Nyakoncept introduceras och svagheter identifieras. Arbetet skall däremotinte anses slutfört utan forskningen fortgår för att kringgå svagheterna,öka på transistorernas stabilitet och framförallt tillämpa dem i innovativaapplikationer.60

ISBN 978-952-12-2468-3Painosalama Oy, Turku, Finland 2010